Intel Haswell CPUs Debut, Put To the Test

Transactional Memory support by rev0lt · 2013-06-01 03:34 · Score: 5, Insightful

For me, this is by far the biggest architectural improvement I see in these line of processors (check http://en.wikipedia.org/wiki/Transactional_Synchronization_Extensions and http://software.intel.com/sites/default/files/m/9/2/3/41604 for more information). If it sticks, it will help solving a lot of multi-core shared memory software development issues.

Re:Transactional Memory support by Z00L00K · 2013-06-01 04:05 · Score: 4, Interesting

It's an interesting addition which can be useful for some.
But when it comes to general performance improvement it's rather disappointing. Looks like they have fine tuned the current architecture without actually adding something that increases the performance at the same rate as we have seen the last decades. To some extent it looks like we have hit a ceiling in increased performance with the current overall computer architecture and that new approaches are needed. The clock frequency is basically the same as for the decade old P4, the number of running cores on a chip seems to be limited too, at least compared to other architectures.
One interesting path for improving performance that may be useful is what Xilinx has done with their Zynq-7000 which combines ARM cores with FPGA, but it will require a change in the way computers are designed.

--
If builders built buildings the way programmers wrote programs, then the first woodpecker would destroy civilization.
Re:Transactional Memory support by PhrostyMcByte · 2013-06-01 05:26 · Score: 4, Interesting

A more immediately useful feature is backwards-compatible hardware lock elision. Before taking a lock, you emit an instruction which is a NOP for older CPUs but causes Haswell to ignore the lock and create an implicit transaction. Instant scalability improvement to just about every app out there with contention, without having to distribute Haswell-specific binaries.
My favorite feature, though, is scatter/gather support for SIMD. Scatter/gather is very important because up until now loading memory from several locations for SIMD use has been a pain in the ass involving costly shuffles and often requires you to load more than you actually immediately wanted, possibly forcing you to spill registers. It's really not something you want to do, but sometimes there are no good alternatives. I'll be super interested to see benchmarks taking this into account.
Re:Transactional Memory support by Hamsterdan · 2013-06-01 05:29 · Score: 2

I'm pretty sure if AMD would bring out something to be competitive, Intel would find a way to come out with faster processors, just like they did when the original Athlon kicked ass. Suddenly they were able to manufactures processors that were'nt just 33Mhz faster than the previous ones.

--
I've got better things to do tonight than die.
Re:Transactional Memory support by cnettel · 2013-06-01 05:32 · Score: 1

But when it comes to general performance improvement it's rather disappointing. Looks like they have fine tuned the current architecture without actually adding something that increases the performance at the same rate as we have seen the last decades. To some extent it looks like we have hit a ceiling in increased performance with the current overall computer architecture and that new approaches are needed. The clock frequency is basically the same as for the decade old P4, the number of running cores on a chip seems to be limited too, at least compared to other architectures.
Even single-threaded performance, even if you normalize for identical frequencies, has increased since Core 2. That they have increased since Pentium 4 goes without saying. We do not see the same increases in instructions per second that we used to do, but we still have increases. This page (and the next) from Anandtech was quite illuminating.
Re:Transactional Memory support by TheRaven64 · 2013-06-01 06:00 · Score: 3, Insightful

The hardware lock elision stuff is going to be more than just a little bit useful. It means that software that uses coarse-grained locking can get the same sort of performance as software using fine-grained locking and close to the performance of software written specifically to support transactional memory. It will be interesting to see if Intel's cross-licensing agreements with other chip makers includes the relevant patents. If it's something that is widely adopted, then it is likely to change how we write parallel software. If not, then it will just make certain categories of code significantly more scalable on Intel than other CPUs.

--
I am TheRaven on Soylent News
Re:Transactional Memory support by Anonymous Coward · 2013-06-01 06:08 · Score: 0

The issue with the market is that ARM marketing and datacenter power consumption point that power reduction is a must, so Intel has to move around trying to improve both performance and power. So you complain about significant performance, when the rest of the world is complaining about power consumption.
Re:Transactional Memory support by bhinesley · 2013-06-01 06:11 · Score: 1

The clock frequency is basically the same as for the decade old P4
The P4 had an absurdly long 31 stage pipeline, compared to the Haswell's 14 stages. More stages means each stage is shorter, and thus more cycles per second. So, sure, the clock frequency is "basically the same", but clock frequency alone doesn't mean a whole lot when you're comparing completely different architectures.
Re:Transactional Memory support by Anonymous Coward · 2013-06-01 06:49 · Score: 1

Instant scalability improvement to just about every app out there with contention, without having to distribute Haswell-specific binaries..
Sorry that's not completely true. Any app that has LOCK contention but little DATA contention will improve. Most highly contended locks have some degree of shared data naturally and if the CPUs collide on said cacheline modifications then the transaction fails. Performance *can* actually suffer. Any contended lock with very little data contention should just be broken up into smaller locks (or some parts not done under the lock).
Re:Transactional Memory support by bored · 2013-06-01 06:52 · Score: 1

You should look at that page again, i'm not sure is benchmarking is "fair". If he wants to compare the older CPU's it might have been ammusing to use the 5 year old version of the code too. That way, your not seeing the affects of code optimized for the latest CPU's at the expense of the old ones.
Ignoring things that are using SSE modes not available on the core2 (cinebench, etc!).
The i7-4770 is actually clocked at 3.9Ghz when running a single core. So for something like 7 zip, 4807/3.9=1232 units/ghz. Vs the core2 at 2519/2.5=1007 units/Ghz. The numbers don't look that great.
We are talking a CPU with more cache, faster RAM, faster PCIe... Its only getting a puny 20% performance improvement in 7-zip single threaded over nearly 5 years!
Now of course that is the pessimistic view, the optimistic is that if you application can take advantage of 256bit AVX, or a dozen cores then the CPU's have gotten significantly faster.
But, I've seen this a few times, someone goes out and purchases a new machine to replace an old high end machine and the new machine is actually slower at some number of given tasks. This happened where I worked, we replaced a couple 3.2Ghz opteron machines with some nice new 2.0 Ghz xeons, and the latency of our application went up ~25% and the throughput went down by ~10%. I'm sure if we had purchased better xeons that wouldn't have happened. But, the guys in purchasing assumed that spending the same amount of money for a new machine as they spent 5 years ago on the old ones would yield a better machine. They were wrong!
Re:Transactional Memory support by Anonymous Coward · 2013-06-01 06:53 · Score: 0

Which is why the wii u is more powerful than people give it credit for.
Re:Transactional Memory support by bored · 2013-06-01 07:01 · Score: 2

Just to reply again, some of these benchmarks are obvious bullshit. Like the AES one, the new cpu's have AES-NI instructions for accelerating AES.
So, yes if you happen to be doing AES, and your running code that can take advantage of AES-NI then the new CPU's are going to fly. But the whole benchmark is so tilted its not even funny. Why not use a benchmark that renders some SSL encoded web pages? Because the benchmark is going to be bottlenecked by the network stack and the rendering engine. Not the tiny percentage of the time the CPU spends doing AES.
He really should have broken the benchmarks into two camps, benchmarks actually running the same code on both sets of CPU's and ones that can leverage some new instructions available only on the new CPUs.
Re:Transactional Memory support by Z00L00K · 2013-06-01 08:02 · Score: 2

Of course the performance has improved since the P4, but the point is that it has been by tweaking things like cache, parallel execution with discard of unwanted branches, pre-fetch, various types of pipelining, out of order execution and so on.
All this means that in order to achieve high performance you have an architecture that does a lot of stuff that eventually is getting thrown away because it was done just in case. The catch here is that it costs energy and builds complexity. The benefit is of course that you can get a processor that executes multiple instructions per clock cycle which improves performance.
However there's a limit on the performance gain these improvements can provide.

--
If builders built buildings the way programmers wrote programs, then the first woodpecker would destroy civilization.
Re:Transactional Memory support by cnettel · 2013-06-01 08:02 · Score: 1

Of course we should use instruction sets that are available. However, I think you need to defend your point on the compiler issue. Most precompiled software is optimized for a pretty old common target. I would be highly surprised if default code emitted by e.g. GCC is slower on a Core now, than it was five years ago. This is also due to the fact that Conroe all the way to Haswell share so many characteristics. Netburst vs. Core entailed far more cases where you would actually prefer differences in emitted code, even for basic things. The relative costs of that pipeline were huge.
Re:Transactional Memory support by edxwelch · 2013-06-01 08:56 · Score: 1

Unfortunately, it's only available on a handful of models. I doubt any software developer is going to use it until it becomes more widepread.
Re:Transactional Memory support by Rockoon · 2013-06-01 09:10 · Score: 1

My favorite feature, though, is scatter/gather support for SIMD. Scatter/gather is very important because up until now loading memory from several locations for SIMD use has been a pain in the ass involving costly shuffles and often requires you to load more than you actually immediately wanted, possibly forcing you to spill registers.
Scattter/Gather will never be as efficient as aligned sequential reads and writes, so all this has bought you is a bit more efficiency on already highly inefficient data layouts.

I see this problem time and again, where a programmer that just doesnt get it is trying to use SSE in sub-optimal ways and then only getting marginal performance improvements over their non-SSE code. It turns into one big waste of time and effort. Essentially, if your data is AoS instead of SoA, then you've already decided that you never want to be highly efficient and layouts that need Scatter/Gather are even worse than AoS layouts.

--
"His name was James Damore."
Re:Transactional Memory support by cnettel · 2013-06-01 09:27 · Score: 2

I wouldn't agree. The out of order depth has increased in later architectures, but not by that much. Rather, branch prediction is better, so we have fewer mis-predicts. We have wider execution units (thus necessitating the higher depth), but that's a "real" improvement in performance. The number of instructions retired per clock under ideal conditions per core, is higher. Hyperthreading has been kept, but it is implemented in a smarter manner with a more dynamic sharing of resources. All of this is happening in the area of "fat" cores and naturally we are reaching some point where there is little more that can be done, but typical energy use under idle and modest load has been decreasing, not increasing. If you just want the performance of a 65 W Core 2 Duo, you can certainly get that out of the new 35 W desktop SKU and get pretty close to it with the 17 W laptop ones under most conditions as well. It is "only" a decrease by 50-75 % in 6 years or so, but it is far from "nothing changes". And, yeah, by the way, you get integrated graphics within that 35 W TDP as well.
Re:Transactional Memory support by gnasher719 · 2013-06-01 10:35 · Score: 4, Interesting

The clock frequency is basically the same as for the decade old P4, the number of running cores on a chip seems to be limited too, at least compared to other architectures.
P4 was an architecture where clock frequency was the design goal, with no regards for actual performance. Any non-trivial operation used lots of cycles, because one cycle was just too short to do useful work. A simple shift instruction was four cycles. An integer multiplication nine cycles. Since Banias, the design goal was performance, not clock speed. The amount of work done in a cycle is vastly increased. The clock speed of P4 and current processors is just not comparable.

On the Macintosh side, Apple shipped pre-release Intel Macs with 3.6 GHz P4s to developers. The first real hardware with 1.83 GHz Core Duos ran _faster_. But if you look at benchmarks, current high-end consumer Macs run about 15 times faster again!

To maybe make a stronger impression: If Apple replaces the processor in the fastest iMac with a Haswell chip, you'll get a computer that would make it into the top 100 of the June 2000 "Top 500 Supercomputer" list. That's how fast a modern Intel computer is, compared to a P4.
Re:Transactional Memory support by Anonymous Coward · 2013-06-01 10:41 · Score: 0

Fortunately things like this are coming into existence.
Re:Transactional Memory support by rrohbeck · 2013-06-01 11:03 · Score: 1

I dearly hope this meager performance improvement gives AMD a chance to catch up a bit.

--
thegodmovie.com - watch it
Re:Transactional Memory support by PhrostyMcByte · 2013-06-01 11:38 · Score: 2

I see this problem time and again, where a programmer that just doesnt get it is trying to use SSE
Do you teach? I don't see that very often outside of SIMD newbies. Curiosities aside, as I said:

It's really not something you want to do, but sometimes there are no good alternatives.
There are some algorithms, like EWA/Jinc image resampling, which simply don't translate well to SIMD without gather support. Keeping with the image processing theme, sRGB gamma decompression will be lightning fast with a L1-friendly 1KiB LUT -- I bet faster than any low-precision approximation in use today.
This is not about trying to fit AoS into a SIMD world. This is about enabling some things in SIMD which simply weren't possible before. Newbies will no doubt use it improperly, but then those with experience and perhaps a little imagination will use it to make code that screams.
Better AoS support shouldn't really be discounted, either -- some standard data formats outside of the programmer's control are really unfriendly to SIMD, and scatter/gather might not give an 8x speedup in processing it but it will certainly help.
Re:Transactional Memory support by PhrostyMcByte · 2013-06-01 11:43 · Score: 1

Why not use a benchmark that renders some SSL encoded web pages? Because the benchmark is going to be bottlenecked by the network stack and the rendering engine. Not the tiny percentage of the time the CPU spends doing AES.
This benchmark is not quite relevant to the average user, but SSL can be a large bottleneck in web servers -- it is certainly relevant there.
Re:Transactional Memory support by gnasher719 · 2013-06-01 11:56 · Score: 1

Let's see if I understand this correctly: If I use a lock to protect an array of data, and two threads access the array simultaneously, but read / write different cache lines of the array, then they will both think they got the lock, but no (time consuming) lock operation has actually happened. But if they access the same cache line of the array, then both threads are restored to the point where they tried to lock, and this time the lock is performed for real on both processors, with one having to wait. And the code to do this will run correctly on older processors, just without the optimisation.
Re:Transactional Memory support by Anonymous Coward · 2013-06-01 13:49 · Score: 0

Not all algorithms can work on aligned data. Like, a tridiagonal matrix multiply--very common operation in scientific apps.
Re:Transactional Memory support by mczak · 2013-06-01 14:36 · Score: 1

That is not quite correct, Haswell (or any cpu which supports AVX2) can do gather but not scatter. I agree this can be useful. I'm not sure how much faster it'll actually be in practice compared to using multiple loads (note that ever since sse4.1 you will typically not need any shuffles to achieve the same with multiple loads as you can use pinsrx instructions to directly fill up those vectors hence your concerns about increased register usage are unfounded), since the instruction is microcoded in the cpu so to the execution core it will still look like multiple single loads.
Re:Transactional Memory support by bored · 2013-06-01 14:41 · Score: 2

The app I work on needs a couple GB/sec (not bit) of encryption bandwidth. So, I love having AES-NI.... But.... We have encryption/compression acceleration hardware in the machine. Which means we won't be using AES-NI because the CPU's haven't increased their compression speeds to match their new found encryption speeds. Plus, we (like many other things, including truecrypt) have always offered the user the ability to encrypt with algorithms that were more CPU friendly than AES.
Anyway, two points. In the high bandwidth web servers I've seen, the SSL public key handshake part of the initial negotiation was the problem with SSL, not the AES/etc. As such, i've seen accelerators for SSL setup too! Frankly, its possible to get ~100MB/sec out of even older CPU's doing software AES. So a 8x CPU machine can pretty much keep a 10Gb link saturated if all your doing is AES... Doesn't leave much time for anything else, but you can either pay for more CPU's or buy accelerators if you actually need the bandwidth. Frankly, adding a couple extra CPU's is not really going to be your problem if your transferring more than 10Gbit of web traffic.
But this is all xeon territory, not desktop CPU's. Which is where truecrypt comes in. If your running it on a recent SSD and you just have to have AES instead of something faster, AES-NI is going to be your thing. But its quite possibly the one desktop application with the largest performance difference between a modern CPU and one a couple generations back. Your not going to be pulling 10x+ performance increases on many other applications. So putting it in a general cpu comparison except as a footnote is a little disingenuous. Same for some of these other things where openCL/CUDA offloads have become the norm so the new AVX/etc instructions aren't necessarily such a big win. Not that I mind having them either!
Re:Transactional Memory support by bored · 2013-06-01 15:07 · Score: 1

Most precompiled software is optimized for a pretty old common target. I would be highly surprised if default code emitted by e.g. GCC is slower on a Core now, than it was five years ago.
Most software is not 1/2 of these benchmarks. I agree you build a generic binary for x86_64 with gcc its basically running the same code on all CPUs. But, povray, cinebench and x264 are all highly optimized. It wouldn't surprise me if all of them are using icc, and have tweaked in on way or the other by intel.
PovRay an open source project has "intel attention" http://software.intel.com/en-us/blogs/2009/03/26/some-tips-for-povray-1
Frankly, this is nothing new. The CPU/GPU/machine vendors have traditionally been willing to help high profile or applications selected for benchmarks run well on their hardware. The problem is that these benchmarks aren't representative of a lot of actual work being done on PCs. If you choose a different video encoding library, you may find completely different results.
Basically, IMHO a significant percentage of the generic performance improvement on intel CPU's in the last 5 years have been benchmarking tricks. Nearly every in house application i've been benchmarking with generic x86_64 gcc targets disagrees with the conclusions drawn on places like tom's hardware. Part of that is gcc (its still not as good as icc, at optimizing for a specific target) part of it is that the basic integer/move/compare/etc instructions that comprise most code haven't gotten much faster. In my tests the only thing intel has done since core2 that affects that kind of code in any significant way has been fixing rep stosd so that it runs at a reasonable rate with respect to the dozen alternative ways to copy memory. This is also why there is a lot of talk that the tables are turned in the AMD/Intel land if your running linux/gcc. Its because AMD looks good and even beats intel in many cases on generic semi optimized linux/gcc code paths that don't leverage SSE4.
Re:Transactional Memory support by smallfries · 2013-06-01 21:17 · Score: 1

Are you assuming that all programmer are working on simple 1:1 transformations of data? It is impossible to encode anything with a summation term without using a gather operation. If there is a projective transformation in the algorithm (i.e. a change of representation onto different axes / number of dimensions) then it is impossible to encode efficiently without a scatter. Perhaps there are more algorithms out there that are suitable for vector architectures that you are familiar with?

--
Slashdot: where don knuth is an idiot because he cant grasp the awesome power of php
Re:Transactional Memory support by rev0lt · 2013-06-02 02:38 · Score: 1

That is, in summary, what HLE does. But RTM requires you to use explicit instructions to signal transactions. Using this approach, you do need code modification and lose backwards compatibility, but gain in granularity and simplicity of implementation.
Re:Transactional Memory support by rev0lt · 2013-06-02 02:42 · Score: 1

Unfortunately, it's only available on a handful of models.
Somehow, I'm not surprised (like the vt/vtx support on core processors). But this is a real improvement for top of the line processors, such as any upcoming xeon products. I'd expect it to be slow in adoption initially, but - in time - to gain traction.
Re:Transactional Memory support by rev0lt · 2013-06-02 03:29 · Score: 1

But when it comes to general performance improvement it's rather disappointing.
If you run benchmarks compiled for previous architectures, it is expected for them to suck - or at least, have the same baseline performance.

To some extent it looks like we have hit a ceiling in increased performance with the current overall computer architecture and that new approaches are needed
He have. To go beyond that, we'll need a completely new technology stack. However, latest trends are short pipelines, specialized instructions, decreasing gate size, and smarter software. They do work quite well, up to a point. If you need beyond that point, you probably can still squeeze some performance by trimming/optimizing the software layer.

The clock frequency is basically the same as for the decade old P4
As it would be expected. However, a single-core - pick your favourite cpu - will run leaps around a P4 at the same speed. One of the biggest differences is the pipeline size, and the stall penalty for cache misses. In many ways, modern Intel CPUs are way more simpler than the P4 line.

(...) at least compared to other architectures
What other architectures? Cores and threads are different things. Some RISC architectures do have simpler implementations, but not with the performance results you want - only more cores per die, and not all tasks are easily paralelizable.

One interesting path for improving performance that may be useful is what Xilinx has done with their Zynq-7000 which combines ARM cores with FPGA
I'd see that as in going the opposite way. FPGA are dead slow for real-life applications, and ARM cores aren't the general purpose computing platform the media has been promising for years. For me, it looks like a prototyping platform to design and test your ASICs, not a consumer-grade platform.
Re:Transactional Memory support by ChrisMaple · 2013-06-02 17:07 · Score: 1

I've seen code that takes different execution paths depending on what machine it's running on.

--
Contribute to civilization: ari.aynrand.org/donate
Re:Transactional Memory support by bored · 2013-06-02 18:48 · Score: 1

That is fairly common when optimizing things.
Especially if your optimizing for different subsets of the instruction set. Trycrypt for sure, is using AES-NI paths on CPUs that support it, otherwise its using generic ones.
My comment was more in regard to gcc, which I don't believe i've ever seen automatically generate multiple arch dependent code paths. I think gcc assumes you will make multiple compilation paths with different -march's and glue them together if you want something better than the generic target.
Re:Transactional Memory support by Anonymous Coward · 2013-06-03 15:31 · Score: 0

P4 was an architecture where clock frequency was the design goal, with no regards for actual performance. Any non-trivial operation used lots of cycles, because one cycle was just too short to do useful work. A simple shift instruction was four cycles. An integer multiplication nine cycles. Since Banias, the design goal was performance, not clock speed.
I'd take issue with the idea that P4 was designed with no regards for performance. Clock speed is a valid way of improving performance. It may not prove to be the optimal way in the long run, but that doesn't mean the designers were deliberately turning their noses up at performance.
The thing which ended P4's viability was that Intel ran into a power wall. P4 was originally a 180nm design, and it got shrunk through three more nodes, 130nm, 90nm, and 65nm. (Some of those "shrinks" were more like "redesigns", but the basic theme of the P4 architecture remained.) As it turns out, these nodes were where the entire chip industry hit a huge problem with increasing leakage current. This problem hit P4 worse than competing lower-clocked CPU designs because you need two very basic things to crank clock rate high: lots of pipeline stages (equals lots more transistors), and very fast transistors (which were hit much worse by increasing leakage than slow transistors). By 90nm and 65nm the P4's shipping clock rates were thermally limited: the chips were capable of hitting far higher clock rates but only at the cost of consuming ridiculous amounts of power.
130nm P4s were actually pretty decent processors for their time.
Re:Transactional Memory support by Anonymous Coward · 2013-06-03 16:30 · Score: 0

I'd see that as in going the opposite way. FPGA are dead slow for real-life applications,
That depends on what you're doing. If you need clock speeds above 300-400 MHz, sure, but if you can refactor for a wide-and-slow datapath instead of narrow-and-fast, there's no reason you can't get FPGAs to support very high throughput on many problems.
These words we're typing now almost certainly pass through FPGAs at some point on their way through the Internet, because it's possible to do this for high data rate networking, and lots of network backbone gear is done in FPGA. Much of what FPGAs are about is exploiting the long tail of the market. Even though an FPGA is always less efficient than an ASIC in terms of die area and power, fixed function ASICs are scary-expensive (and risky) to design, test, and manufacture. By making one chip which can serve thousands of different tiny niches, the FPGA vendors get huge economies of scale, and their customers get to avoid the up-front costs (and risks) involved in ASIC design.

and ARM cores aren't the general purpose computing platform the media has been promising for years. For me, it looks like a prototyping platform to design and test your ASICs, not a consumer-grade platform.
No, not true at all. Unlike Xilinx's old FPGAs that had PowerPC cores, Zynq isn't a FPGA with a processor added, it's an ARM SoC with FPGA fabric added. It has all the usual hard IP blocks expected of a generic Cortex-A9 SoC -- memory controller, USB, GbE, etc. The FPGA fabric is there to let you add your own functionality to the SoC, usually some kind of application accelerator. It's a classic FPGA vendor long tail strategy -- they're making it possible for low volume systems houses to design their own ARM SoC without all the cost and risk of actually spinning an ASIC. (And without the cost and risk of doing their own Linux port. Don't discount this, it's a huge effort. It's much easier to add drivers for your one accelerator peripheral to the pre-baked Zynq Linux.)
As for prototyping, well, I work on prototyping ASICs in FPGA at a fabless semiconductor company. We evaluated Zynq, and it's not a good match. The reason why is that you can't turn off the SoC stuff! Ideally you want access to the raw Cortex-A9 interface, just like you get in your own ASIC. In Zynq, the A9 sits behind Xilinx's SoC infrastructure. You're stuck with the Zynq memory map, Zynq SoC peripherals, interrupt controller, and so forth. You get no access to the Cortex-A9's cache-coherent ACP bus interface either. If your ASIC needs only a tiny chunk of the memory map and has no need for cache coherence, you could probably prototype with Zynq, but for us it was no good.
Re:Transactional Memory support by Anonymous Coward · 2013-06-04 09:15 · Score: 0

That does not follow from what bhinesley said. He's just pointing out that you have to actually measure performance instead of blindly trusting it to scale with clock speed.
There are plenty of reasons for anyone familiar with CPU architecture and the history of CPU designs to expect that the Wii U's CPU probably isn't all that powerful. It's an evolution of the Wii's CPU, which itself was an evolution of the Gamecube CPU. The Gamecube's was basically an IBM PowerPC 750cx with a few added features. The 750 (aka G3) was a cutting edge architecture once -- in 1997 or so. It's kinda not, these days. I'm sure the latest version in the Wii U is considerably better, but Nintendo hasn't exactly been pursuing high CPU performance in the console generations prior to the Wii U and I doubt this strategy has significantly changed.

Performance per Watt by Technician · 2013-06-01 03:36 · Score: 4, Insightful

Hmm, Performance per Watt seems to have been glazed over.

The possibility of a fanless media center PC, the ability of a server farm to eliminate over half the cooling cost, and long battery life in a gaming class laptop seems to not be the attention of the article.

Gee it's only X percent faster..

--
The truth shall set you free!

Re:Performance per Watt by gl4ss · 2013-06-01 03:37 · Score: 1

Hmm, Performance per Watt seems to have been glazed over.
The possibility of a fanless media center PC, the ability of a server farm to eliminate over half the cooling cost, and long battery life in a gaming class laptop seems to not be the attention of the article.
Gee it's only X percent faster..
still uses plenty of juice when gaming..

--
world was created 5 seconds before this post as it is.
Re:Performance per Watt by Anonymous Coward · 2013-06-01 03:53 · Score: 0

I don't think ANYONE in his right mind will do ANYTHING even remotely CPU intensive on a laptop that is NOT plugged in.
Performance per Watt is relevant only for situations requiring either very large (supercomputers) or very small (appliances) performance. Gaming for instance is not one of these.
Re:Performance per Watt by Anonymous Coward · 2013-06-01 03:56 · Score: 0

The post is referring to the power drain while under load, which the quote in the summary talks about. "single largest improvement in battery life in Intel history" is quite the claim to not support.
Re:Performance per Watt by Anonymous Coward · 2013-06-01 04:16 · Score: 0

Speak for yourself. The main reason why PCs are still not the main media centre / gaming unit in households is due to the ridiculous noise they make. PSU has a fan, GPU has a fan, CPU has a fan, chassis has 2 to 8 fans, some MBs have small fans. Even going water cooling means you have bloody annoying pumps.
Using less Wattage means less cooling required. A little above you perhaps?
Re:Performance per Watt by Gadget_Guy · 2013-06-01 04:56 · Score: 1

I don't think ANYONE in his right mind will do ANYTHING even remotely CPU intensive on a laptop that is NOT plugged in.
What nonsense. Gaming capable notebook computers are quite affordable these days. I have one to play games on my long train trips to work. Not only do more efficient computers mean that we can play on batteries for longer, but they also stop our genitals being fried from CPU/GPUs under heavy load.
Similarly, my main gaming PC is pretty low spec solely because I don't want to have a furnace sitting at my feet, especially in summer when I used to have to stop playing games when it got too hot with my previous system. More efficient processors mean that I don't have to have such a weak system just because I want cool & quiet gaming.
Re:Performance per Watt by hpa · 2013-06-01 04:59 · Score: 2

I do kernel hacking on airplanes. 'nuff said.
Re:Performance per Watt by Molochi · 2013-06-01 05:35 · Score: 1

Yeah they only tested the power consumption of the new unlocked/enthusiast desktop CPU (4770k) with a 84W TDP.
Toward the beginning of the article they say that there will be a 35W i5 Haswell for socket 1150. No mention on specs tho'

--
"The Adobe Updater must update itself before it can check for updates. Would you like to update the Adobe Updater now?"
Re:Performance per Watt by aliquis · 2013-06-01 06:42 · Score: 1

our genitals being fried from CPU/GPUs under heavy load.
There must be a joke in there waiting to be found but I don't know what it is :)
Re:Performance per Watt by Kjella · 2013-06-01 08:29 · Score: 3, Interesting

Anandtech tested it, idle power is down probably due to the new voltage regulator (FIVR) but active power.... 113% the performance for 111.8% so performance per watt is essentially unchanged. If what you need is CPU power then you're better off waiting for a IVB-E hex-core in Q3, in threaded applications a quad-core Haswell won't touch a hex-core Ivy Brigde - it's trailing Sandy Bridge hex-cores as well. If you're not interested in the graphics or battery life, it's a giant yawn.
That said, the GT3e graphics for mobile looks to carve out a solid niche in the notebook market, the R-series desktop processors (GT3e graphics, BGA only) is probably compelling for AIOs that don't have room for graphics upgrades anyway and the lower idle wattage should be good for all laptops with Haswell graphics. None of the processors launched now have the new idle states for ultramobile/tablets, so the effect of those we'll have to wait to see. Anandtech tested the i7-4950HQ and it was impressive how a 47W mobile chip consistently beat AMDs A10-5800 100W desktop APU in gaming benchmarks. Of course it's going to sell in a price range of its own, but AMD just lost the crown here.
As a CPU in a regular tower with discrete graphics it's at best incremental but I think the full launch lineup hit all of Intel's main competitors - it's threatening AMD and nVidia's low end discrete card sales, it's threatening AMDs APU sales and the lower idle power is promising for their lower power parts that will compete with ARM. They're just not winning much against the i7-3770K but then they're also fighting against themselves in that market, the FX-8350 is not even close. The 8-series chipset finally brings 6 SATA3 ports, so the main AMD advantage chipset-wise also disappeared.

--
Live today, because you never know what tomorrow brings
Re:Performance per Watt by Anonymous Coward · 2013-06-01 08:38 · Score: 0

Tablets were the last bastion of fanless computing
No, phones are the last bastion of fanless computing.
Re:Performance per Watt by Seumas · 2013-06-01 08:59 · Score: 1

Couldn't have said it better, myself. A year after Ivy Bridge and a big null for most non-corporate purchasers of Haswell. At ten percent or less annual improvement, it's going to take you many years (or some phenomenal fundamental change in architecture that revolutionizes everything) to make a new CPU compelling.
Re:Performance per Watt by 0123456 · 2013-06-01 11:14 · Score: 1

Speak for yourself. The main reason why PCs are still not the main media centre / gaming unit in households is due to the ridiculous noise they make.
Maybe if you buy AMD's space heaters, but my i7 + GTX 660 system is barely audible when playing games.
You must be confusing modern PCs with Pentium-4s or something.
Re:Performance per Watt by 0123456 · 2013-06-01 11:19 · Score: 2

Fact is, the emphasis is still on processing power over performance per watt. Sure, Intel gives it lip service, but not much beyond that.
So why does my new i7 system use half as much power as my old Pentium-4?
Re:Performance per Watt by Anonymous Coward · 2013-06-01 14:36 · Score: 0

Partly because Intel have been improving efficiency over the years, and partly because the Pentium 4 was fundamentally a dud.
Re:Performance per Watt by Technician · 2013-06-02 03:26 · Score: 1

I have a fanless laptop. The only noise is the hard drive. When I can afford a SSD (when my HD dies), I'll go silent.
Low power enables small form factor PC's. Consider a Media Center PC that simply attaches to the back of your flatscreen TV..

--
The truth shall set you free!

Progress by Anonymous Coward · 2013-06-01 03:38 · Score: 1

Intel's Haswell architecture would likely be the most amazing thing to happen to the tech industry in years.

Seems I've been hearing this about each of their "tocks".... Nehalem, then Sandy Bridge, and now again with Haswell.

Their process is good, and that kind of advertising may even be warranted, but the hype is really getting repetitive.

Re:Progress by ArcadeMan · 2013-06-01 03:57 · Score: 2

The hype is really getting repetitive.

--
Get free satoshi (Bitcoin) and Dogecoins
Re:Progress by Samantha+Wright · 2013-06-01 04:58 · Score: 2

The hype may be the most amazing thing to happen to the tech industry in years.

--
Bio questions? Ask me to start a Q&A journal. Computer analogies available for most topics!

Software killed the PC, not hardware by Anonymous Coward · 2013-06-01 03:41 · Score: 4, Insightful

The lack of phenomenal hardware improvements may annoy the nerds, but the mass market PC is killed by the abysmal software environment. People are fleeing to tablets and phones and with that the cloud because maintaining a PC has become just about impossible for laymen. The slowness of a desktop that hasn't seen professional maintenance is astonishing, if it is still working at all. Viruses and trojans aside, every bit of software comes with its own updater, many of which are poorly implemented resource and attention hogs. If the updater doesn't do you in, it's the bundled adware, sometimes installed by the update "service". The PC is stiff and stone cold, a host overwhelmed and killed by its parasites. Time to put it 6ft underground.

Re: Software killed the PC, not hardware by Anonymous Coward · 2013-06-01 04:12 · Score: 0

Couldn't agree more. This needs to be upvoted to 5 and then given more for extra measure. PC's days on the Desktop are numbered. When mobile platforms have their GUIs optimised for the Desktop, that will be it for PCs. Only Desktops needed for specialist apps (that don't have a non-PC equivalent) and Servers will survive.
Servers will be next though. Soon we will have hundreds of tiny ARM based systems setup in a software cluster, in the same physical space as 1 x86 based server, achieving a fraction of the power usage but with powerformance that is magnitudes better.
Re:Software killed the PC, not hardware by Anonymous Coward · 2013-06-01 04:20 · Score: 1

Absolute bollocks. There's nothing special about tablets that makes them immune. What? Exchange the power to install what I want for a crApp Store? No thanks. Fuck the cloud too. I'm backing down all my data from Flickr after their latest fiasco. Paying the play in somebody's walled garden is no better than paying a subscription fee for AV that slows down your machine. Long live the PC. Fuck the tablets. People will get sick of paying for stuff that should be free. People will get sick of the cloud when it costs them money, and not just nerds. I know a "layman" who lost a cloud app when the OS updated because the stupid thing didn't keep track of the fact that he had paid for it. Their only "solution" was for him to pay again. Shit like that will kill the cloud of tablets (sounds like a bad acid trip), and good riddance.
Re:Software killed the PC, not hardware by Anonymous Coward · 2013-06-01 04:59 · Score: 0

oh wow, some ignorant ass cant keep is P4 XP machine running cause they have not installed a single update on the thing since it came from the factory!
dude, stop living in 2005
Re:Software killed the PC, not hardware by MightyMartian · 2013-06-01 05:16 · Score: 1

You forgot to add "Now get off my lawn!"

--
The world's burning. Moped Jesus spotted on I50. Details at 11.
Re:Software killed the PC, not hardware by TheRaven64 · 2013-06-01 06:03 · Score: 1

There's nothing special about tablets, except for one thing: users don't expect backwards compatibility. You can make significant improvements if your customers don't expect to be able to use your product as a drop-in replacement for something else.

--
I am TheRaven on Soylent News
Re:Software killed the PC, not hardware by PopeRatzo · 2013-06-01 06:31 · Score: 4, Interesting

Time to put it 6ft underground.
This was the giveaway.
What do you care if there are still people who would rather use a desktop PC that's not behind a garden wall and actually get work done? Why do you insist that the PC platform has to be killed off? Isn't there room in the world for more than one set of computing needs?
This notion, that only the most popular form of anything should exist pops up strangely often around here. The iPad is phenomenal, so Android tablets should just disappear from the market. The iPhone is popular so no Windows phones can be allowed. That sort of thing.
Friend, I can understand that you'd rather work on a tablet and have someone else make decisions about what you can and cannot have, what you can and cannot do, but why in the world are you so insistent that no one else be able to make their own choice?
I don't get you.

--
You are welcome on my lawn.
Re:Software killed the PC, not hardware by etash · 2013-06-01 06:38 · Score: 2

do you know how many people have declared the PC to be dead. It's usually either people who have "better" solutions to offer, or the useful idiots who believe them. Ever tried doing some real work on a tablet ? like video editing, image editing, mundane tasks like excel, word editing. how about video games ?
Re:Software killed the PC, not hardware by drinkypoo · 2013-06-01 06:58 · Score: 1

The workstation isn't going to die any time soon, but it is being marginalized by game consoles one the one hand and portable devices on the other. I perform more and more tasks on my phone because it is close to hand and because there's an app for that. I can now reasonably get some information on an address faster by firing up maps on my phone (which is a 2011 model, and not exactly hot shit) than by walking into the next room, or look up a wikipedia article on the same basis.
The hobbyist computer is being replaced by inexpensive SBCs, so only diehard gamers and various professionals actually need big ugly desktop computers any more. I've still got one and I'm not planning on giving it up any time soon, but if some non-technical type asked me what they should buy, I'd ask them what they wanted to do with it and probably wind up recommending a tablet, or a phone upgrade and some STB-class computer, game console, or literally a STB.

--
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Re:Software killed the PC, not hardware by lightknight · 2013-06-01 07:15 · Score: 1

You're right. The PC industry needs to come out with maintenance free PCs, for the common people!
We will take the automotive industry's example of oil-free cars, and run with it. No longer will you need to stop by Good Year / Jiffy Lube / etc. or (shh, is forbidden) change your oil yourself: this is the year 2013, right, WTF do they mean read the car manual and perform maintenance? Don't they know that no one has time for that? Just make cars / trucks / boats / planes maintenance-free, and your problem is solved! Besides, those car mechanics probably steal from you when you bring your car in...or break something so that you need to bring it back in...and $40 for an oil change? It's a rip-off!
Hell, we should replace the keyboard with 'The Facepalmer (TM),' because the computers these days are smarter than the end users, so they should know what they want to do, right? It's not that hard! Just have the user smash their face into 'The Facepalmer (TM),' and it will reorganize their iTunes collection / send an email for them / or browse the web for them! It's that easy!

--
I am John Hurt.
Re:Software killed the PC, not hardware by Seumas · 2013-06-01 09:05 · Score: 2

The 40 or 50 some-odd million Steam gamers and the often 6 million concurrent online Steam gamers would beg to differ with you and, anecdotally, the new people I hear every day saying that they're building a PC gaming rig for the first time in their life is surprising and seems to be growing. There are a ton of gamers out there and they have always been a heavy push for the PC market. They aren't going anywhere. They continue to be just as important and just as numerous, even in a world full of consoles and tablets and cell phones.
Also, most people who would ONLY use a phone or a tablet are not people who would have bought a desktop, anyway. They MIGHT have otherwise had a low power laptop... MAYBE... but they were never going to be users going out and building/buying meaningful desktop systems. These aren't losses from one platform fleeing to another.
Re:Software killed the PC, not hardware by Seumas · 2013-06-01 09:09 · Score: 2

I don't buy the "consoles are killing PCs" argument for one second (also, people have spent the last two years claiming that tablets and phones are killing consoles, so there you go).
Steam *alone* has something like 40 or 50 million users and between 2.5 and 6 million concurrent users playing a game at any one moment.
We don't even need to talk about the depth and breadth of games available on PC that simply don't exist on other platforms. Just with the user numbers alone -- and only the Steam ones, here -- compare those to consoles and you simply can not argue that PC games are dying. I think that a lot of factions (including publishers, themselves) are doing every fucking thing they can to invalidate, complicate, and ultimately kill PC gaming and that has been the case for a few years . . . but it isn't changing the fact that it's pretty fucking vibrant and charging forward.
Re: Software killed the PC, not hardware by vlueboy · 2013-06-01 09:25 · Score: 1

About allowing both sides to exist, I understand there is a need for diversity but it is generally only appreciated by the geek community... and even here, people always secretly play favorites. Given that anything that can go wrong will do so, we end up with garbage decisions being incorporated by both sides little by little. Being broad here: think loss of 4:3 screens/resolution, crappy A/V software now coming to mobile OSs, App stores now unavoidable on windows and mac desktops, mobile GUIs taking over PCs to the point your old program is suddenly cryptic and useless (think MSN Messenger shoving us to the Metro only Skype on Windows 8 a month ago), and fans not going away... let me add bufferbloat delays now at your cable box and TV, and annoying LED lights from everything in your bedroom at 4am. So I dont mean killing PCs or tablets is warranted, but pollution potential is there. Did I mention the idea of freemium Apps could not exist without the former learning exp of shareware and nagware of old? It is mainstream now and will not go away, creating a world where free desktop applications will grow smaller each year
Re:Software killed the PC, not hardware by drinkypoo · 2013-06-01 10:02 · Score: 1

I don't buy the "consoles are killing PCs" argument for one second (also, people have spent the last two years claiming that tablets and phones are killing consoles, so there you go).
That's not the argument.

Steam *alone* has something like 40 or 50 million users and between 2.5 and 6 million concurrent users playing a game at any one moment.
And a number of them will probably buy a Steambox in the not too far future instead of upgrading their PC.

We don't even need to talk about the depth and breadth of games available on PC that simply don't exist on other platforms.
Personal computers have not gone away. They are simply diversifying.

--
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Re:Software killed the PC, not hardware by im_thatoneguy · 2013-06-01 11:06 · Score: 1

Depends on your definition of a "tablet". I've been using a Surface Pro as both a tablet and a PC for the last few months and love it. Editing 5k RED footage, creating rough composites, recording naration, excel, word editing and playing games. And then I just pop the keyboard and use it as a tablet for web browsing on the couch.
Re: Software killed the PC, not hardware by Anonymous Coward · 2013-06-01 11:13 · Score: 0

^ Annoying slashdot mobile crippled my post: No car returns for some reason, at least in plain text...
I'm on the desktop now and can see what a textwall that was!
While I'm on the subject, what's the deal with slashdot redirecting to m.slashdot without a Desktop option? The link says it will take you there, but just redirects back to m.slashdot. Not friendly at all for my 2 year old single core Froyo phone.
Re: Software killed the PC, not hardware by Anonymous Coward · 2013-06-01 12:22 · Score: 0

Does your tablet run Dwarf Fortress?
Re:Software killed the PC, not hardware by Anonymous Coward · 2013-06-01 13:46 · Score: 0

Are you sure you're talking about PCs? I think you're talking about one particular type of sub-standard one. Most OSes aren't anything at all like what you just described.
Re: Software killed the PC, not hardware by PopeRatzo · 2013-06-01 13:59 · Score: 1

I guess if you play Skyrim, or COD, or any other PC game you are considered part of the "geek community" because those are the only people left who care about the desktop platform.
Forget Photoshop, video production, music production, graphic arts, development, writing, etc etc, we're just the geek community and we better get on board with a tablet because the desktop needs to be put "6ft under" according to our friend.
I wonder if he thinks all those marvelous apps on his phone were created on an iPad.

--
You are welcome on my lawn.
Re:Software killed the PC, not hardware by Anonymous Coward · 2013-06-01 14:42 · Score: 0

The lack of phenomenal hardware improvements may annoy the nerds, but the mass market PC is killed by the abysmal software environment.

Absolute bollocks. There's nothing special about tablets that makes them immune.
You mean other than the software environment? Tablets essentially rebooted the software environment and have lower end CPUs which requires the new software ecosystem to be more careful. You can certainly argue that it's less capable (and I'd agree) but it's definitely less wasteful right now.
Re:Software killed the PC, not hardware by Anonymous Coward · 2013-06-02 05:00 · Score: 0

Hey bro, where do you think the crApp Store its based, runned and backed? In another tablet...?
Come on...this remind me the day where we were arguing about why windows its ways better, or mac or linux.
Just take what you like, or need or can.
the macs arent for everyone, the pcs arent for everyone, the tablets arent for everyone.
You make all the work in your mobile: great!!!, you are enough with your tablet: marvelous, you are still using w95, you are right!!!
Just live and let it live.
Re:Software killed the PC, not hardware by GuB-42 · 2013-06-03 01:04 · Score: 1

The PC is not dead.
Sales are low simply because people already have a PC that is good enough for them. A PC bought 5 years ago, when iOS and Android didn't even exist, is still usable.

How does this compare by maroberts · 2013-06-01 03:52 · Score: 4, Interesting

With AMDs CPU/GPU solutions?

--

Donte Alistair Anderson Roberts - hi son!
Karma: Chameleon

Re:How does this compare by Anonymous Coward · 2013-06-01 04:01 · Score: 1

It should probably have said 'In graphics, Haswell completely redefines the expectations for Intel processor graphics.'
Re:How does this compare by Anonymous Coward · 2013-06-01 04:04 · Score: 0

Same thing just slightly more advanced and more expensive
Re:How does this compare by Anonymous Coward · 2013-06-01 04:05 · Score: 0

CPU: Intel's still way ahead for anything constrained by Amdahl's law, AMD's probably a bit faster when something can actually use all those cores.
GPU: Intel's winning for once, but that may not hold up under examination of the price ranges. AMD puts their good GPUs in low-end CPUs, Intel puts them in high-end CPUs. If Intel hasn't mixed that up with this release too much (I haven't investigated), AMD may still be ahead in practice.
Power/thermals: If you care, you shouldn't even be considering AMD.
Re:How does this compare by Reliable+Windmill · 2013-06-01 04:24 · Score: 0

You can't factor in external graphics boards for this comparison, and AMD A10 still beats the new Intel on graphics.

--
Signature intentionally left blank.
Re:How does this compare by Anonymous Coward · 2013-06-01 04:54 · Score: 0

According to Anandtech's benchmarking at least, the new Intel stuff looks to be solidly ahead of the A10. What I was referring to is that the current Haswell chips cost quite a lot more than an A10.
If I were trying to build a decent all-around rig, not specifically for gaming, I wouldn't be looking at the i7's price range. More like an A10. At the AMD A's price range, Intel has historically had worthless GPUs. They've only been stepping that up in higher end stuff, but if I'm paying enough for a CPU that Intel's throwing in decent graphics, I'm going to want a real GPU to go with it, rendering it a moot point.
Re:How does this compare by Osgeld · 2013-06-01 05:00 · Score: 1

meaning its almost not embarrassing
Re:How does this compare by Reliable+Windmill · 2013-06-01 05:17 · Score: 1

It's in part 11 - IGP Performance: AMD A10 (Radeon HD 7660D) is up to 20% ahead of the new Intel (Intel HD 4600) in graphics performance. When you factor in external graphics boards, or tests which are CPU-bound, you get different results.

--
Signature intentionally left blank.
Re:How does this compare by Molochi · 2013-06-01 05:48 · Score: 1

They compared the 4770k to the A10. A10 was still faster in games. 4770K was faster in OpenCL
Tom's also said intel will have a faster integrated graphics setup (iris pro) but it will be exclusive to the BGA offerings and not offered on LGA1150 CPUs.

--
"The Adobe Updater must update itself before it can check for updates. Would you like to update the Adobe Updater now?"
Re:How does this compare by beelsebob · 2013-06-01 05:55 · Score: 2

Nope, the Iris 5200 wipes the floor with the A10. You're probably referring to the low end HD 4600 in the i7 (which they put there because they expect no one buying an i7 to be using integrated graphics).
Re:How does this compare by beelsebob · 2013-06-01 05:56 · Score: 1

The point being that the HD 4600 is the slowest Haswell IGP. Look at the Iris 5200 instead on anandtech's benchmarks. It's a good 20-30% ahead of the A10's 7660D.
Re:How does this compare by Anonymous Coward · 2013-06-01 07:11 · Score: 1

It's also only in $400+ mobile i7s.
Re:How does this compare by beelsebob · 2013-06-01 08:58 · Score: 1

Right, in reality intel is offering no chip that directly competes with the A10. They're currently offering the i3 3220 (ivy bridge) which is a faster CPU, but has a slower GPU. I'd imagine that there'll be a haswell variant along some time soon that's both a faster CPU, and roughly equal GPU, but it's not out right this second.
Re:How does this compare by Rockoon · 2013-06-01 09:18 · Score: 4, Interesting

It's also only in $400+ mobile i7s.
Intel fanboys never want to discuss price, at least not really. Sometimes they pretend, but never want to make an honest price bin comparison.

--
"His name was James Damore."
Re:How does this compare by Anonymous Coward · 2013-06-01 09:27 · Score: 0

It should be noted that Intel's Iris Pro isn't slated for any of the slower LGA1150 CPUs either.
Re:How does this compare by Kjella · 2013-06-01 10:30 · Score: 2

Intel fanboys never want to discuss price, at least not really. Sometimes they pretend, but never want to make an honest price bin comparison.
So let's compare single threaded CPU performance at equal price points then shall we. No? Ah right, the comparison of choice lately has been gaming performance at equal price points. Choose the scenario where you win and pretend it's the only scenario that matters, it's a powerful form of self-delusion. But that aside, they don't compete directly just yet but Intel just proved they have an integrated graphics solution that is faster and uses less power than anything AMD can come up with. That is what most people would call "technical superiority" where you can extract tons of profit by offering products your competitors can't. Intel has a lot of that, AMD just lost one of the last strong cards in their deck.
Take a grounded reality check - is AMD's pricing because they're this undiscovered treasure chest of excellent processors or because they're selling inferior solutions they won't sell any other way? I think you can certainly rule out that it's a strategic plan from AMD's side, given how their latest financial statements look. Intel is not pricing Haswell aggressively at all, if anything they're increasing prices so that maybe AMD can manage to not go bankrupt and get Intel in all sorts of scrutiny over their x86 monopoly. And still AMD is failing, really if they can't manage to undersell Intel at the prices Intel have now maybe AMD doesn't deserve to be in business. Sad for the consumer, but true.

--
Live today, because you never know what tomorrow brings
Re:How does this compare by beelsebob · 2013-06-01 11:27 · Score: 1

AMD fanboys never want to discuss the fact that the i7 4770 is roughly 4 times faster than the A10 The only reason they're compared is because the A10 is the fastest chip with an IGP AMD ships, and the i7 is the slowest haswell out just now. By the time the i3 4220 lands, the A10 will be dead in the water.
Re:How does this compare by Rockoon · 2013-06-01 14:26 · Score: 0

So let's compare single threaded CPU performance at equal price points then shall we. No?
Yes. Lets do that.

Ah right, the comparison of choice lately has been gaming performance at equal price points.
Oh, I thought you were going to do that thing you said in the sentence before. But you didn't, as I predicted. You failed to make an equal price comparison even after suggesting that you would.

And no, this second sentence is not true either. For example, the previous comparison article on slashdot compared a $600 Intel system with a $300 AMD system.

Now are you going to make the comparison or not?

--
"His name was James Damore."
Re:How does this compare by Rockoon · 2013-06-01 14:37 · Score: 0

AMD fanboys never want to discuss the fact that the i7 4770 is roughly 4 times faster than the A10

Wrong. We discuss it. But when we start talking about price you people go back to this. Exactly this.

The i7-4770 costs 4+ times as much (currently if you pre-order) as the most expensive AMD A10 chip, which also arent even the fastest AMD chips available.

Lets see what you did:

Price was brought up, so you responded by not mentioning price at all, and proceeded to compare the highest performing Intel chip ever with a mid-range AMD part.

I'm glad that you shit so much money that its not a factor in your buying decisions. Is that a Macintosh you've got there?

--
"His name was James Damore."
Re:How does this compare by beelsebob · 2013-06-01 15:06 · Score: 4, Insightful

Wrong. We discuss it. But when we start talking about price you people go back to this. Exactly this.
The i7-4770 costs 4+ times as much (currently if you pre-order) as the most expensive AMD A10 chip, which also arent even the fastest AMD chips available.
i7 4770k –$339
A10 5800k - $129.99
So we're looking at a chip that's 2.6 times the price, but 4 times the CPU performance.

Price was brought up, so you responded by not mentioning price at all, and proceeded to compare the highest performing Intel chip ever with a mid-range AMD part.
No, the assertion that the intel chip was 4 times more expensive was brought up, and neglected to mention the context that the intel chip was also 4 times faster (and also greatly exaggerated the 2.6 times price). I merely noted that there's a reason for the intel chip being significantly more expensive – that it's significantly faster.
You're right, the AMD chip is indeed a mid range chip, and it should never even have been compared to the i7. Unfortunately, AMD has no chip that can reasonably be compared to the i7, so all the review sites chose the fastest CPU AMD has on offer that includes an IGP. If you want to push it a bit further, lets ignore the IGP completely, and take the fastest desktop chip AMD has on offer:
FX 8350 –$199.99
i7 4770k –$339
So, we're at 1.9 times the price, and around 1.8 times the performance (see here).
But, of course, we all know that price increases exponentially with performance when it comes to high end processors, so lets just scale back that intel part a bit:
FX 8350 –$199.99
i5 4670 –$213
So now we're at 1.06 times the price (effectively equal), but about 1.5 times the performance (see here).
If you really want to make the price war completely in intel's favour, then look at
FX 8350 –$199.99
i5 4570 - $192
Unfortunately Anand do not list this in their benchmarks, but it's the exact same die as the 4670, just clocked 6% lower, so we can extrapolate the result, and say that this i5 will be around 1.4 times the speed of the FX 8350 Oh, and by the way, it throws in a decent IGP, while the FX 8350 does not. It also consumes only 67% of the power under load, and that's including the IGP.
So conclusion. Haswell is pretty effectively blowing AMD out of the water. The A10 5800 remains reasonably unscathed only because Haswell i3s have not yet been released. When the i3 4220 appears, I would fully expect the A10 to lose to it in every respect in the same way as the FX 8350 loses to the i5 3570 in pretty much all ways.
Re:How does this compare by tyrione · 2013-06-01 16:42 · Score: 1

Nope, the Iris 5200 wipes the floor with the A10. You're probably referring to the low end HD 4600 in the i7 (which they put there because they expect no one buying an i7 to be using integrated graphics).
I look forward to you eating it when the APUs based on Steamroller waltz in and slap the shit out of this overpriced heater.
Re:How does this compare by beelsebob · 2013-06-02 04:19 · Score: 1

This "overpriced heater" that uses half the power of the A10-5800k, right?

Graphics.. by ottawanker · 2013-06-01 03:58 · Score: 0

Haswell completely redefines the expectations for processor graphics.

.. for Intel. It still seems to lag behind AMD's on-die graphics.

Re:Graphics.. by timeOday · 2013-06-01 04:10 · Score: 2, Informative

You must have said that before looking at the benchmarks? Looks to me like AMD is toast. Intel's integrated graphics beat AMD's on every game in the AnandTech test.
The new Intel even beats the discrete mobile GPU (Geforce GT 650M) on a couple tests. On most the Intel is somewhat slower but using around half the power.
Re:Graphics.. by Reliable+Windmill · 2013-06-01 04:22 · Score: 0

You've misinterpreted the benchmarks, or read the wrong ones. Look at 11 - IGP performance: the AMD A10 still beats the new Intel in graphics performance, and it does so at half the price.

--
Signature intentionally left blank.
Re:Graphics.. by illaqueate · 2013-06-01 04:50 · Score: 1

that said the only article I saw there was comparing a $650 chip vs a $130 AMD A10 chip. and the only game it's close to the 650 in performance is Crysis Warhead which is heavily CPU limited. in other games they were only competitive where the settings were low which not only allowed the Intel card to close performance but also the A10.
the 4770k which is the $339 part is slower than the a10 in games according to this
http://www.techpowerup.com/reviews/Intel/Core_i7_4770K_Haswell_GPU/
Re:Graphics.. by timeOday · 2013-06-01 05:11 · Score: 1

The AnandTech review I was referring to is the HD 5200 integrated graphics in the mobile chip. The Techspot one you are referring to is for desktops where Intel has not included HD 5200, it is the HD 4600, so the AMD won by 7% - 24% for integrated graphics performance. I guess the HD 5200 is not coming in desktop CPUs for a few more months, but it is about 50% faster than HD 4600 (both are on the AnandTech charts) so I hope AMD has something big up its sleeve.
As to the price difference, CPU performance has to count for something too. On media encoding the Intel is more than twice the speed of that particular AMD. And with a discrete graphics card the Intel beat every AMD in the test. And at idle the Intel takes less than half the power of an AMD FX system.
Re:Graphics.. by beelsebob · 2013-06-01 05:57 · Score: 1

On the other hand, the Intel CPU was 4-5 times faster than the A10 on every CPU bound benchmark ;)
Re:Graphics.. by Kjella · 2013-06-01 08:59 · Score: 1

That said the only article I saw there was comparing a $650 chip vs a $130 AMD A10 chip.
So far, because that chip:
1) Trounces AMD in CPU performance
2) Uses less than half the power
3) Is a fully featured mobile chip
where
4) Is somewhat faster in graphics
is only a small detail.
The interesting chip here is the i7-4770R, OEM only and no reviews yet but the same graphics will come to the desktop somehow. I expect it will be sold as motherboards with CPU soldered on for DIY builds, like the Atom boards.

--
Live today, because you never know what tomorrow brings
Re:Graphics.. by edxwelch · 2013-06-01 09:06 · Score: 1

The GT3e only is available on 3 48W TDP models and at a very high price. AMD doesn't make any mobile APU at this TDP or price. What will be more interesting is a comparison between GT3 and Richland.
Re:Graphics.. by Sloppy · 2013-06-01 15:37 · Score: 1

The interesting chip here is the i7-4770R .. I expect it will be sold as motherboards with CPU soldered on for DIY builds, like the Atom boards.
I sure hope so. Maybe this is projection but I think such a mobo would be insanely popular for desktops.

--
As copyright owner of this comment, I authorize everyone to defeat any technological measure which limits access to it.

How appropriate by ArcadeMan · 2013-06-01 03:58 · Score: 0, Offtopic

The WWDC is only 9 days away.

--
Get free satoshi (Bitcoin) and Dogecoins

Re:How appropriate by Anonymous Coward · 2013-06-01 05:27 · Score: 0

Everything always has to be about YOU doesn't it.
Re:How appropriate by dbraden · 2013-06-01 08:14 · Score: 1

That was my first thought, as well. Just in time for a new Apple product release.

Anand Lal Shimpi by sartwell · 2013-06-01 04:00 · Score: 1

Anand wrote that summary, not jjslash.

Re: Anand Lal Shimpi by Anonymous Coward · 2013-06-01 04:06 · Score: 0

Technically, jjslash put everything Anand wrote in quotes, so the summary is a quote of a quote.
Re: Anand Lal Shimpi by Anonymous Coward · 2013-06-01 05:56 · Score: 0

Package mine to take out. And add a double order of wontons too.

Re:Optimized for Macbook Air by Anonymous Coward · 2013-06-01 04:07 · Score: 0, Insightful

You'll need the electricity savings to offset the absurd cost of the MacBook Air.

If it saves you $50 in electricity per month over your present system, all other factors being equal, in 3 and a half years you'll be saving money. People don't buy luxury to save money. Relevant car analogy, not a single 2013 Ferrari model gets 20 MPG or better. By contrast, most Kia vehicles get 30 MPG or better.

The moral of this story? Poor people care about economy, wealthy people care about performance, and business people look for a happy medium.

Mod parent up by Anonymous Coward · 2013-06-01 04:09 · Score: 0

first paragraph

Need to wait a few years by gnasher719 · 2013-06-01 04:13 · Score: 4, Insightful

In the last few years, Intel has been adding new instructions that will give major performance gains when they are used. For example, Haswell can do two fused multiply-adds with four double or eight single precision operands per cycle per core, but no current code will use this. We'll get the advantage when HPC code is recompiled (in a few months time), and when general code assumes that everyone has this feature (in five years time). But on the other hand, we _now_ get the advantages of features they added five years ago.

Re:Need to wait a few years by Anonymous Coward · 2013-06-01 04:30 · Score: 3, Informative

This is absolutely the most overlooked aspect of Haswell. As an incremental improvement, its less than stellar, but in certain areas, it effectively doubles performance over the previous generation. Aside from FMA for floating point apps, the integer 256bit SIMD pipeline is now effectively feature complete.
Your point about waiting for recompiles is a good one - all the more reason we should be moving to adaptable metaprogramming systems for HPC, rather than a constant manual reworking of codebases. Projects like Terra (http://terralang.org) are particularly promising in this regard
Re:Need to wait a few years by Anonymous Coward · 2013-06-01 08:52 · Score: 0

Surely out of order execution can take advantage of this in some cases even without a recompile?

Overclocking potential by pellik · 2013-06-01 04:19 · Score: 1

I don't know what the OP is talking about saying it has only the same overclocking potential.

The -K series has unlocked the base clock multiplier allowing it to be set up to 166MHz without destabilizing the other controllers on the die. This should allow for considerably more fine-tuning.

Also, the theoretical maximum overclock is 8GHz (80x100MHz, 64x125MHz, or 48x166MHz). 4.6GHz may still be a reasonable goal for an air cooled system, but there is certainly more potential.

Re:Overclocking potential by Anonymous Coward · 2013-06-01 04:31 · Score: 0

Not to mention 4,5GHz and above overclocks on low voltages versus what you'd need on Sandy/Ivy.

Heat Dissapation by stms · 2013-06-01 04:21 · Score: 2

How hot do these chips get. I have to throttle my (not overclocked) 3770k when I do media encodes because it gets too damn hot. I've been meaning to get a better cooler I just haven't gotten around to it.

Re:Heat Dissapation by Anonymous Coward · 2013-06-01 05:57 · Score: 1

There's no point in throttling the CPU manually. It'll automatically start throttling itself when it gets too hot (105C or higher). The thermal shutdown temp is 130C. Nevertheless, if a non-overclocked 3770k keeps hitting 105C, you might want to check that the cooler is seated correctly and that the thermal paste has been applied properly.
Re:Heat Dissapation by DaveGod · 2013-06-01 06:48 · Score: 1

Hmm, does the Intel-supplied fan make quite a racket when it starts getting hot? When it gets hot the fan should speed up considerably compared to idle. Also the system should be throttling the CPU automatically if it gets too hot. If these are not happening, I suggest checking your BIOS settings (I assume you are not running any tweaking software supplied by the manufacturer, which is usually very clunky). Another possibility is the hot air is not being exhausted.
If you end up getting a new cooler, have a look at some of the excellent reviews of the Coolermaster Hyper 212+, which is very popular, inexpensive and I can confirm the reviews as regards it being effective and quiet.
Re:Heat Dissapation by stms · 2013-06-01 07:16 · Score: 1

I'm testing it right now and I'm getting 98C as my max which is a bit more than I'm comfortable with. I think I may have seen it a bit higher before. This is the first rig I've put together by myself so it's possible I may have done something wrong. I've heard that the 3770 runs hot that's why I bring it up.
Re:Heat Dissapation by stms · 2013-06-01 07:17 · Score: 1

It gets a bit louder not to the point that it's too bothersome. My mobo is a GIGABYTE GA-Z77X-UD5H all I did to it when I built my system is upgraded the firmware and configure boot options. Thanks for the tip on the the cooler I was considering the Noctua NH-D14 which is a bit more pricey but if it keeps my chip from frying it'll be worth it.
Re:Heat Dissapation by Rockoon · 2013-06-01 09:28 · Score: 1

There's no point in throttling the CPU manually. It'll automatically start throttling itself when it gets too hot (105C or higher).

WHATS THAT? I CAN"T HEAR YOU OVER THE FAN. CAN YOU PLEASE REPEAT YOURSELF?

The reason people underclock is temperature, but the symptom they are trying to eliminate is the noise.

--
"His name was James Damore."
Re:Heat Dissapation by Anonymous Coward · 2013-06-01 09:41 · Score: 0

I have a 3770k @ 4.2GHz with a coolermaster 212 evo. Idle proc is 38C and when encoding video it is 52C
Re:Heat Dissapation by Kjella · 2013-06-01 10:59 · Score: 1

Sounds like you should check your fan, my CPU was hitting 99C under load and the reason was that the cheap plastic attaching it to the motherboard had failed from fatigue and the fan was loose on one side but still marginally attached to the processor, providing enough cooling for light loads. No properly attached cooler should have a problem with a 77W processor, not even the cheap OEM ones. The reason I noticed was because it was getting slightly unstable, so since I didn't want to do any more thermal damage to it I got a big 80mm push-pull fan that keeps it very cool. Seems to work so far.

--
Live today, because you never know what tomorrow brings
Re:Heat Dissapation by Anonymous Coward · 2013-06-02 17:16 · Score: 0

How hot do these chips get. I have to throttle my (not overclocked) 3770k when I do media encodes because it gets too damn hot. I've been meaning to get a better cooler I just haven't gotten around to it.
Get around to it!
Something like a Noctua NH-D14 (for example, there are lots of choices) will let you run the CPU flat out without throttling.
Heck, that cooler will let you run a hex-core Sandy Bridge E flat out without throttling, and that's less efficient than a 3770K.

AMD still a winner by Reliable+Windmill · 2013-06-01 04:28 · Score: 0

Intel do have the big, raw CPU power, but I think that's not something the majority of desktop users care about. AMD A10 still beats the new Intel on graphics, which is probably not key to most desktop users, but the price is: 4 out of 5 desktop users have all their CPU and GPU needs covered in the AMD A-series, at half or even one third the price of the equivalent Intel. This is the key: if you go AMD you save lots of money.

--
Signature intentionally left blank.

Re:AMD still a winner by Anonymous Coward · 2013-06-01 04:36 · Score: 0

Wel, depends of course, since you get more per wat, for datacenters there is pretty much no competition. I heard somewhere that for anything that requires big processing power, you should pretty much replace everything every 2 years and will save money (if its for the same processing power of course).
Re:AMD still a winner by Anonymous Coward · 2013-06-01 07:41 · Score: 0

Wel, depends of course, since you get more per wat, for datacenters there is pretty much no competition. I heard somewhere that for anything that requires big processing power
It is not power per watt. It is performance per dollar that matters. That includes hardware, software, and power (for server and cooling). Yes, generally after 2 or 3 years, replacing at least the CPU is well worth it, unless you are running idle most of the time.
For home or office machines, performance per watt is meaningless unless you are running your CPU 100% for some crazy reason. For home or office, it is idle power and capital costs that matter most. If a CPU is fast enough, who cares what the performance per watt or concurrent threads?
I have an AMD CPU. CPU utilization is 1.5%. It is fast enough. Cost me $100 2 years ago. Spending another $100 for better performance per watt or more computing power would have been a complete waste. On the other hand, spending $50 extra on a high efficiency power supply is well justified.

Mobile Graphics by Daniel+Hoffmann · 2013-06-01 04:34 · Score: 1

It seems the graphics for the notebook chips are pretty good: http://www.notebookcheck.net/Intel-HD-Graphics-4600.86106.0.html One step above nVidia 540m (the one I have on my laptop) http://www.notebookcheck.net/Mobile-Graphics-Cards-Benchmark-List.844.0.html For reference this could play recent games on medium settings very well. For gamers that can't afford (or don't want to bother with) two computers it seems this chip will provide enough horse power to play most games, a thrown back to the era (486 dos days) where you could simply play all available pc games if your rig was not older than 2 years. These days you have to actually buy a dedicated (and expensive) gaming pc to play games, I think this is what gave the console makers the edge on the market. Intel can change that.

Why would Intel care by zrelativity · 2013-06-01 04:49 · Score: 4, Insightful

Why would Intel care about raw CPU performance. They have no competition from AMD in CPU performance. The GPU performance may not be as good as A10, but it has improved and that's what matters for Intel.

Intel for a little while has correctly perceived that their risk to business is from shift in computing to mobile devices and they are addressing this issue. One thing Intel has always been very good at, and I'm a great admirer of them for that, when they perceive a risk, they are extremely good at steering their giant ship very rapidly into the headwind and tackle that threat. Their process technology lead also gives them a huge advantage.

Over the next couple of years, the battle front will be the mobile and server devices, the desktop processors will become a second class citizen. Maybe this will give some lifeline to AMD, but AMD is so far behind on performance.

Re:Why would Intel care by Samantha+Wright · 2013-06-01 05:02 · Score: 1

A monopoly must always strive to slightly outdo itself, so that it may motivate its captive market to continue consuming. Regardless of the technical challenges inherent in improving performance now, a 10% improvement really says "here's something we can force the next cycle of bleeding-edge adopters to take up."

--
Bio questions? Ask me to start a Q&A journal. Computer analogies available for most topics!
Re:Why would Intel care by The+End+Of+Days · 2013-06-01 07:09 · Score: 0

I'm not sure you understand the meaning of the word "force"
Re: Why would Intel care by cerberusss · 2013-06-01 10:04 · Score: 0

LOL, they completely and utterly missed the smartphone and tablet market. The iPhone and iPad should have contained something intel, but it didn't.

--
8 of 13 people found this answer helpful. Did you?
Re:Why would Intel care by Anonymous Coward · 2013-06-01 14:17 · Score: 0

Why would Intel care about raw CPU performance. They have no competition from AMD in CPU performance. The GPU performance may not be as good as A10, but it has improved and that's what matters for Intel.
What?
The GPU performance of the new high-end Iris Pro 5200 blows away the AMD A10. The mobile i7-4950 varies between 110% and 150% of the desktop A10-5800K... when AMD releases a mobile A10 for benchmark, it won't look pretty. Source: AnandTech: Intel Iris Pro 5200 Graphics Review: Core i7-4950HQ
Re:Why would Intel care by Anonymous Coward · 2013-06-01 17:47 · Score: 0

Again ignoring the fact that the intel part is $600 and AMD part is $130... sigh..
Re: Why would Intel care by Anonymous Coward · 2013-06-01 18:01 · Score: 0

They will in less than 2 years
Re:Why would Intel care by Anonymous Coward · 2013-06-02 10:41 · Score: 0

Over the next couple of years, the battle front will be the mobile and server devices, the desktop processors will become a second class citizen.
No, in the worst case, desktops will simply end up with the low-end server CPUs, which isn't actually all that bad. We've seen that pattern before.
- T

Get A Clue, Intel by Jane+Q.+Public · 2013-06-01 04:55 · Score: 4, Insightful

While I am all for advances in CPUs, I seriously wish Intel would go back to a naming scheme for its CPUs that made any kind of sense to the average buyer (or even the technically-oriented buyer). I have grown really weary of having to look at tables of CPU specifications every time I shop around for computers.

Intel's naming scheme -- expecially in recent years -- has been a mishmash of names and numbers without any obvious coherence. Get a clue Intel. You're hurting your own market.

If I didn't have to run OS X in my business, I'd buy AMD just for that reason. Their desktop CPUs may not be quite up to the latest Intel, but they are certainly adequate and the price is better.

Re:Get A Clue, Intel by XanC · 2013-06-01 05:57 · Score: 1

I thought people who said "expecially" and "excape" just had sloppy pronunciation. Is it possible that they actually think that's how those words are spelled??
Re:Get A Clue, Intel by spire3661 · 2013-06-01 06:01 · Score: 1

http://ark.intel.com/

--
Good-bye
Re:Get A Clue, Intel by Anonymous Coward · 2013-06-01 06:18 · Score: 1

Fat fingers. S and X are right next to each other on a QWERTY keyboard.
Re:Get A Clue, Intel by theskipper · 2013-06-01 08:25 · Score: 1

Racist but pretty clever. +1 Funny
Re:Get A Clue, Intel by Anonymous Coward · 2013-06-01 10:17 · Score: 0

expecially in recent year's

FTFY.
Re:Get A Clue, Intel by Jane+Q.+Public · 2013-06-02 12:45 · Score: 1

It was just a slip of the finger. "X" is right under "S".

amd also has more MB choice and more pci-e lanes by Joe_Dragon · 2013-06-01 04:58 · Score: 1

amd also has more MB choice and more pci-e lanes.

With lntel you need Pci-e switchers on the MB to get more then 16 pci-e lanes (not counting the 4+DMI for chip set link).

Re:Long way to go by Junta · 2013-06-01 05:03 · Score: 1

I'd consider the fact that the most demanding android applications are arm specific in terms of compile is the more critical thing.

Intel does have a product with high core count, Phi has 50 cores for example.

I haven't seen a lot of evidence that ARM under load offered better price performance than Intel before. The only thing making that claim I can recall committed a grevious mistake, measuring ARM power usage and then assuming TDP value rather than measuring x86 power usage. It was undeniable that under typical smartphone/tablet conditions Intel did horribly (mostly 'idle' but immediately able to do things on demand), and that seems to have been the engineering focus this time, a shallower sleep state that's possible without screen blanking is one notable facet.

While Intel's prospects in the mobile arena are slim (Andorid has a lot of momentum, and that momentum is largely tied to ARM in much the same way as Windows is tied to x86), I suspect they will continue to rule the roost in the datacenter and workstation-like workloads.

--
XML is like violence. If it doesn't solve the problem, use more.

*Some* codebase will need rework. by Junta · 2013-06-01 05:08 · Score: 1

For the vast majority of even HPC code, it means compiler rework and math library development. The vast majority of benefit can be achieved with a drop in of a new library without rebuild of the application. In your example, it would be the interpreter, which actually tend to be the last things to receive this attention.

--
XML is like violence. If it doesn't solve the problem, use more.

Re:*Some* codebase will need rework. by Anonymous Coward · 2013-06-01 05:40 · Score: 0

Terra compiles against current LLVM builds; LLVM has supported AVX2 (in some form) since 2011

Re:AMD still a LOSER for me by fnj · 2013-06-01 05:12 · Score: 2

For me. I couldn't care less what the majority of desktop users care about. What I care about is a CPU with high performance per watt and graphics good enough to scroll a text display (editor, eclipse) with some graphics (browser) fast enough not to be annoying, and to watch HD video without any pauses. Sandy bridge was more than good enough in the graphics department for anything I would ever want to do on current displays. Haswell will probably more than maintain this on the highest resolution displays coming down the road in the next few years.

Dollar cheapness of the CPU within limits does not impress me at all. I am generally content with a system for at least 4 years to spread out that cost, and there are many other cost centers in a system besides CPU.

I have never used a single AMD system and see no reason to believe that they will ever make anything that would change my mind.

I don't really begrudge any slack jawed gamers their massive nuclear power plant busting AMD systems with the absurd overkill of space heater SLI graphics; it's just not anything that has the slightest relevance to anything I could ever care about.

Re: Long way to go by LDAPMAN · 2013-06-01 05:19 · Score: 1

For desktop/laptop use, idle power is what matters.

Re:Optimized for Macbook Air by PiSkyHi · 2013-06-01 05:53 · Score: 1

It's extending battery life, not money saving.

Re:Optimized for Macbook Air by Anonymous Coward · 2013-06-01 05:53 · Score: 0

And then there's no-poor-but-not-rich-either people who just want a system that fucking works.

benchmarks vs benchmarks by Anonymous Coward · 2013-06-01 06:21 · Score: 0

Maybe he was looking at the THG test (4770K vs A10) where Intel lost on every game.

Haswell is NOT faster than Ivybridge by Anonymous Coward · 2013-06-01 06:22 · Score: 0

Every tech site knows Haswell has completely failed to improve over Ivybridge, but they are hiding this fact in the 'small print'. The Ivybridge and Haswell chips that Intel calls 'equivalent' show a small advantage to Haswell, but look at the power consumption of the two parts. The Haswell part uses more power than the so-called equivalent Ivybridge when running the same code, and gets a smaller performance boost over Ivybridge than its increased power usage.

Modern Intel parts do NOT run a simple constant speeds under the hood. Intel's stated base and turbo clocks are a gross simplification, and do not compare generation to generation. In other words, the Ivybridge and Haswell parts Intel states are equivalent are NOT equivalent. The Haswell part does more work at default by clocking higher on average (thus its much higher power usage under load).

At a similar power usage, Haswell seems to be somewhat slower than Ivybridge- proving another massive Intel failure in its FinFET project. Intel had attempted to redesign its transistor technology after the massive disappointment of Ivybridge FinFET performance. The truth is that FinFET concepts are under-performing horribly (for everyone) versus the early hype for this new approach to the transistor. At the same time, the vastly simpler/cheaper FD-SOI technologies are showing incredible potential at current process sizes.

FinFET has given Intel 'knee' improvements where power consumption just doesn't matter that much- namely at the low end of mains powered desktop use (or notebooks with whopping great batteries, and limited usage life per charge). On the other hand, high performance main-powered parts have got worse, and ultra-low power mobile parts have barely improved- Intel's worst nightmare.

AMD's Steamroller cores are due to catch up with Intel's cores very soon now, and AMD is going to massively out-pace Intel on RAM bandwidth (by going to GDDR5 and then a 256-bit memory interface as with the PS4). AMD is also converting to a fully unified memory architecture (already present on the PS4 APU from AMD) by the end of 2014, and will have a mostly unified part (Kaveri) at the end of 2013.

Even Intel is switching to new memory technologies with its next generation CPU, making the Haswell an expensive dead-end for ill-informed suckers. Intel is stuck between a rock and a hard-place with the coming focus on integrated graphics for both games and high performance FPU calculations. While Intel can and does improve its own dreadful integrated graphics system, it always lags massively behind AMD and Nvidia, while producing solutions that are vastly more expensive than either of its competitors.

The mobile Haswell with embedded L4 cache for the GPU is a great example of this. A mega-expensive solution from Intel with a fraction of the games performance of an equivalent solution from either Nvidia or AMD/ATI. What notebook manufacturer would be mad enough to use this part? All the expense, and none of the performance of a mid-end gaming notebook using GPUs from AMD or Nvidia.

Brief summary by JDG1980 · 2013-06-01 06:51 · Score: 1

TL;DR: Haswell is OK on the desktop, but nothing special; roughly 5%-10% better than Ivy Bridge. If you're on Sandy Bridge or better already, it's probably not worth upgrading. This architecture was designed for laptops first and foremost. Light power consumption/TDP on mobile parts, and a better GPU, are the big selling points. Apple will get much better integrated graphics so they don't need a Nvidia chip for their top rMBP, but they'll pay out the nose for it.

This is what all the rumors and leaks over the past couple months said and it's now officially confirmed.

Re:Brief summary by Anonymous Coward · 2013-06-01 07:10 · Score: 0

TL;DR: Haswell is OK on the desktop, but nothing special; roughly 5%-10% better than Ivy Bridge. If you're on Sandy Bridge or better already, it's probably not worth upgrading. This architecture was designed for laptops first and foremost. Light power consumption/TDP on mobile parts, and a better GPU, are the big selling points. Apple will get much better integrated graphics so they don't need a Nvidia chip for their top rMBP, but they'll pay out the nose for it.
This is what all the rumors and leaks over the past couple months said and it's now officially confirmed.
You mean "the consumers" will pay out the nose for it. ^_^
Re:Brief summary by Seumas · 2013-06-01 09:18 · Score: 1

The days of significant performance improvement from one year to the next seem to have long since passed. I no longer factor the CPU into my upgrades (well, I don't upgrade - I just build new rigs). Now, it's all about the GPU. When NVIDIA or ATI come out with the next big cards that I just have to get my hands on, that's when it is time to build a new rig and that's when it is time to just get whatever the best CPU out at the time is.
I can't remember the last time when the CPU performance over my current CPU performance was ever a factor in my decision that it was time to build a new rig or in what to put in it.
It's nice that they have something to offer in mobile devices, but I think geeks are primarily interested in the CPU itself as a thing they can use rather than how it might perform later this year in a mass produced device that someone else is going to build that they may or may not eventually buy for themselves. I kind of miss the "this is for the guy at home" marketed CPUs.

Re:AMD still a LOSER for me by drinkypoo · 2013-06-01 06:54 · Score: 1

Dollar cheapness of the CPU within limits does not impress me at all. I am generally content with a system for at least 4 years to spread out that cost, and there are many other cost centers in a system besides CPU.

It's nice to have the kind of money where you can just throw it away, but the majority of people on this planet do not have that kind of luxury.

I don't really begrudge any slack jawed gamers their massive nuclear power plant busting AMD systems with the absurd overkill of space heater SLI graphics; it's just not anything that has the slightest relevance to anything I could ever care about.

I don't really begrudge any slack-jawed gamer-haters their massive pocketbook-busting intel processors with the absurd price tag, but the fact is that the majority of users' needs are covered better by a cheap AMD chip than an expensive intel chip, because their needs include low price.

--
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"

Re:AMD still a LOSER for me by Anonymous Coward · 2013-06-01 07:06 · Score: 0

What I care about is a CPU with high performance per watt

I have never used a single AMD system

Either you're lying or you used a P3 until mid 2006...

Need more cores, 8, 10, 12... by Anonymous Coward · 2013-06-01 07:17 · Score: 0

Why waste the silicon on a GPU that is just mediocre? Yes, you can replace a discrete video card with it, but that does not need to be 1/10th as good as this.

Only on some Haswells, not on "K" processors by niceworkthere · 2013-06-01 08:33 · Score: 2

For some yet unknown reason the unlocked "K" Haswells (and maybe others) were so far all listed to come without TSX.

Most of the noise in my system by msobkow · 2013-06-01 08:45 · Score: 1

Most of the noise in my old P4 based system is from the power supply and the video card. Not that I game any more, but you need 3D acceleration to run the latest and greatest desktops under Linux.

The lower power draw of Haswell and the more-than-adequate 3D support would fit my needs just about perfectly. That 10% performance boost that the article snears at would mean the machine would only run roughly 16 times as fast as my current box.

I think I could live with that. A quiet, low-power system that just does what I need: encode and play back videos, run database software, Eclipse, and day to day browsing/email.

--
I do not fail; I succeed at finding out what does not work.

Re:Optimized for Macbook Air by flargleblarg · 2013-06-01 09:00 · Score: 1

A MacBook Air in typical use uses 10 to 20 Watts, but let's say for the sake of discussion that it uses the full 45 Watts of its power adapter. If you left it plugged in 24x7 for a month, drawing that much electricity, you would use about 33 kilowatts, or about $0.50 in electricity. So your figure of "If it saves you $50 in electricity per month" is off by a factor of 100 to 300.

Re:Optimized for Macbook Air by Anonymous Coward · 2013-06-01 09:26 · Score: 0

you would use about 33 kilowatts, or about $0.50 in electricity

I want your electric rates! Where I live, 33kWh costs over $4.

How does Haswell work with a Graphics Card by Anonymous Coward · 2013-06-01 09:32 · Score: 0

If I get a Haswell system and a Nvida (or AMD) graphics card will the graphics on the Haswell chip work with the card to improve performance? Or will the Haswell graphics basically be useless at that point?

Re:Optimized for Macbook Air by KZigurs · 2013-06-01 11:24 · Score: 1

Show me something comparable* to MacBook Air and I'll be more than happy to try it out.

*Price. Quality. Resale value. Longevity.

Apple have found a very interesting niche mass-producing high-end quality products. Nobody else has managed to come close quality wise and all that investment in taking a BSD kernel and sprucing it up for modern expectations has paid off quite nicely too. The cost is not absurd, in fact, for what I'm getting, I consider it surprisingly cheap.

Re:AMD still a LOSER for me by Smauler · 2013-06-01 12:11 · Score: 1

I have never used a single AMD system and see no reason to believe that they will ever make anything that would change my mind.

So... you'd continue to buy Intel even if someone else made a better processor? A few years back, AMD did have better processors... better in both performance/price and performance/watt. I had one, and it was a great PC which I ran for 7 years or so... now I'm back with Intel. I really don't understand blind brand loyalty.

I don't really begrudge any slack jawed gamers their massive nuclear power plant busting AMD systems with the absurd overkill of space heater SLI graphics; it's just not anything that has the slightest relevance to anything I could ever care about.

High end gaming systems are almost invariably Intel based at the moment, because they are the quickest processors on the market at the moment. Low / mid gaming systems are more often AMD, because they offer as good or better performance / price. Lower end gaming systems are unlikely to run SLI configurations.

The slack jaw gamers you're spouting drivel about (all the time claiming you don't care about) don't exist.... if they do, they run Intel.

Re:Optimized for Macbook Air by Anonymous Coward · 2013-06-01 13:33 · Score: 0

The cost is not absurd, in fact, for what I'm getting, I consider it surprisingly cheap.

That is because you are in the target demographic. A Ferrari owner feels the same way about their shiny new Ferrari, and neither the hypothetical Ferrari owner nor yourself are incorrect. For another demographic, my father, a 2003 era Windows XP machine that cost $600 and is still running today has been an outstanding value.

If all you care about are "Price. Quality. Resale value. Longevity." then a good machine is arbitrarily defined based on your opinion and relative financial status. I would personally never buy anything based on it's resale value because I do not resell things. I buy low cost, use it til it breaks, and then fix it until it's "more weld than original metal", to borrow a line from my dad. To me the word quality means I can repair it myself if it takes physical damage. From my perspective a Macbook Air is just about the least desirable laptop ever, but there again is the "target demographic". I'm not in it.

Apple makes a great product for people that like it. Ferrari make nice cars for the people that buy them. Microsoft makes a nice operating system for a lot of people, and Kia makes good economy cars for the masses. When it comes to perceptions of value, all you have to go on are opinions and biases, both personal and cultural. Windows vs Mac is like Jesus vs Muhammad. By that I mean that Linux is the only true operating system.

Even a Nintendo DS has more than one core by dbIII · 2013-06-01 16:26 · Score: 0

So let's compare single threaded CPU performance at equal price points then shall we

The match looks like it's over, but wait folks, in this exciting development Kjella has moved the goalposts right back to the single CPU heyday of 1999! He scores by moving the posts over the ball! The crowd goes ... oh wait, they've wandered off and don't care. That's it, Kjella wins but nobody cares because it's his own personal little game.

Meanwhile back in reality there are different products for different niches instead of a one size fits all device. In some areas AMD has nothing, in others Intel has nothing and pretending otherwise is utterly ridiculous.

Depends on your field by Frobnicator · 2013-06-01 16:52 · Score: 4, Informative

If all you care about is the perspective of the boring desktop business app, then this processor doesn't have much to excite you. Of course, that's just one field. Sending a few database queries over the wire or updating your text boxes doesn't exactly saturate a quad-core box. Business desktop apps don't really see much no matter what.

For data-heavy, cache-intensive, and parallel-intensive programs the processor looks to offer quite a lot. HPC developers like that.

For notebooks and low-power devices the processor is wonderful. If you are paying the power bill for a data center, the energy use will add up. Accountants and laptop users will like that.

The option to have graphics integrated to the chip this way means better SOC options. Embedded developers will like that.

Many fields will see great things out of this chip.

If you are fixated on the world of desktop business software, you still get an incremental ~10% improvement. Unlike technologies such as SIMD, you get it without changing a line of code. So now you can add 10% more text boxes to fill out, or maybe pick up some more wasteful coding habits.

--
//TODO: Think of witty sig statement

Re:Optimized for Macbook Air by KingMotley · 2013-06-01 19:08 · Score: 1

Then you'd love mine. About $0.065/kWh in the Chicago suburbs. But that's what you get when you actually use nuclear power (55% of our power comes from nuclear reactors).

intel by Anonymous Coward · 2013-06-01 23:01 · Score: 0

also part fo the group that wants to put DRM on software and hardware to control you

Re:Optimized for Macbook Air by Anonymous Coward · 2013-06-02 00:36 · Score: 0

And even your rates would cost $2.15 for 33kW of electricity. The GP's 50 cents just isn't real.

Re:AMD still a LOSER for me by rev0lt · 2013-06-02 04:26 · Score: 1

A few years back, AMD did have better processors... better in both performance/price and performance/watt.

They were better on specific tasks, not consistently better. And those "cheaper" AMD cpus tend to forget to shutdown due to overheat, so it was quite easy to kill one on a 24/7 desktop operation. Besides proper 64 bit support (back when no one was using it :D), those cpus had nothing to offer except a cheaper price and a non-refundable policy. Try to return a fried processor to the supplier, and see how lucky you are.
Then, add to that the fact that most chipsets made for AMD cpus were absolute crappy. I'd go any day with an Intel mobo, instead of an AMD roulette even if it means I need to fork more money for it. Its not about loyalism, its about consistency. I use computers to make money, they're a tool. I don't care if - to buy something more reliable (even if it performs worse) I need to pay more.

On battery efficiency based solely on TDP by espinozahg · 2013-06-02 05:55 · Score: 1

Let us consider 3rd and 4th gen Intel Mobile Processors with 4 physical cores. As you can infere, we are focusing on energy efficiency for high performance mobile processors.
The TDP (Thermal Design Power) for 3rd and 4th gen
3rd gen (Ivy Bridge)
http://en.wikipedia.org/wiki/Ivy_Bridge_(microarchitecture)#Mobile_processors
35 to 55 Watts

4th gen (Haswell)
http://en.wikipedia.org/wiki/Haswell_(microarchitecture)#Mobile_processors
47 to 57 Watts (probably 37 Watt versions will appear)

I cannot see the "huge energy efficiency improvement". Maybe Haswell changes from high power state to low power state faster than 3rd Gen (which is good indeed) but that does not matter to the eyes of my examination.

Re:On battery efficiency based solely on TDP by Anonymous Coward · 2013-06-04 10:54 · Score: 0

Ivy and Haswell TDPs aren't directly comparable because Haswell integrates voltage regulation circuitry into the processor itself, which means some of the system's power budget has moved from the motherboard into the CPU. The system as a whole is more efficient.
Also, the efficiency gains in Haswell aren't just about the integrated voltage regulator permitting quick power state changes. In some versions, the fast voltage ramps and some other new design elements permit Haswell's idle state to fully power down large parts of the core, resulting in about the same power usage as the sleep states in Ivy Bridge. When you see talk of 50% better battery life, this is what they're talking about. It'll probably apply mostly to ultrabooks, since this feature also depends on other parts of the system having very tight integration with the CPU's power management.
It's also worth nothing that the Ultrabook class CPUs actually had their TDP drop. Sandy/Ivy ULV CPUs were 17W, Haswell ULV is 15W (and there's even one 11.5W model). The thing you're not accounting for at all is that TDP in isolation isn't efficiency. Efficiency is a combination of two factors, power used while the CPU's idle (which is pure waste) and compute power per watt while running real code (when the processor actually dissipates its thermal design power). If a new processor core delivers better compute per watt, designers can choose to drop TDP while trying to keep compute power approximately constant, or keep TDP the same while gaining compute power, or even increase TDP and gain even more compute. Intel's choices about how to use Haswell's compute efficiency weren't uniform across all Haswell product lines.
Finally, the 37W versions are real. Wikipedia's out of date. Here's the complete list:
http://ark.intel.com/products/codename/42174/Haswell

but that does not matter to the eyes of my examination.
Are you a non-native English speaker? If so, FYI, that phrase is extremely clumsy in English. It sounds like a translated idiom which just doesn't work well in English. "That doesn't matter in my eyes" would be a far better choice.

Re:Optimized for Macbook Air by Anonymous Coward · 2013-06-04 10:07 · Score: 0

You'll need the electricity savings to offset the absurd cost of the MacBook Air.

Somehow nobody mentioned one of the prime bits of stupidity here: your proof of absurd pricing is a link to some random bozo on Amazon Marketplace who's listed an out-of-production 2011 model (Sandy Bridge CPU) for about $500 more than you'd pay for a brand new 2012 Air (Ivy Bridge CPU) direct from Apple's online store. (And oh by the way, speaking of Haswell, wait a few weeks and the 2013 Haswell model will probably be out.)

After clicking around a bit on amazon, it looks like they've become a terrible place to buy a new Mac. They used to stock a wide variety of new machines themselves but they're barely doing it any more. Almost everything you get when searching is used or new old stock, and most of it's being sold by non-Amazon sellers.

Slashdot Mirror

Intel Haswell CPUs Debut, Put To the Test

189 comments