Slashdot Mirror

You are reading incorrectly then. The link you post is a comparison of i7 7700K with and with two different builds of Windows. Presumably one with the fix and one without. The 5-50% degradation is between old processors before i7-7700k with and without the fix.

The issue is that old Intel processors don't have the PCIDs which help mitigate most but not all of the performance penalty incurred by this fix. So companies with large install base of older Xeon processors such as v2 series Xeons are more likely to experience the slowdown than a company with a brand new server with the latest generation of Intel CPUs.

i7-7700K supports PCID
http://www.cpu-world.com/cgi-b...

Same for my desktop, an AMD A8-3860, which was introduced in 07/2011. Does what I need, quickly enough.

Are these machines high end machines now? Absolutely not... However, the time of buying new toys just to have new toys, is over for me. Works for me, means: no reason to upgrade. Many people who are not into tech think that way. A few years ago, I helped a non-tech with her old desktop. It had died: caps gone up in smoke. I said: hey, it's about 5 years old, it had a good run. She: *only* five years? Non-tech people think differently (Ha!) These days I'd be pissed too if my machine died after 5 years of use.

Same for my desktop, an AMD A8-3860, which was introduced in 07/2011. Does what I need, quickly enough.

Are these machines high end machines now? Absolutely not... However, the time of buying new toys just to have new toys, is over for me. Works for me, means: no reason to upgrade. Many people who are not into tech think that way. A few years ago, I helped a non-tech with her old desktop. It had died: caps gone up in smoke. I said: hey, it's about 5 years old, it had a good run. She: *only* five years? Non-tech people think differently (Ha!) These days I'd be pissed too if my machine died after 5 years of use.

Not sure if stupidity comes naturally to you, but comparing a $950 server cpu to a budget $100 desktop cpu is a little moronic to say the least.

Next thing you'll complain about is how your Xeon is more powerful than a ARM Cortex!
Oh, and while in the thread, don't forget to compare a browser developed by a multi-billion dollar corporate spying empire to one developed by a miniscule non-profit chartable organisation; like so many others have the audacity to do so here.

Not sure if stupidity comes naturally to you, but comparing a $950 server cpu to a budget $100 desktop cpu is a little moronic to say the least.

Next thing you'll complain about is how your Xeon is more powerful than a ARM Cortex!
Oh, and while in the thread, don't forget to compare a browser developed by a multi-billion dollar corporate spying empire to one developed by a miniscule non-profit chartable organisation; like so many others have the audacity to do so here.

The 920 slightly overclocked

If you're overclocked, that changes things. Most people don't do that, I'm comparing stock clock to stock clock.

The i3-6100 can be overclocked as well, with the right ASRock motherboard and BIOS version, but I don't count that as reasonable.

http://www.cpu-world.com/Compa...

The i3-6100 is 64% faster in single (and dual) threaded applications. Only applications that can properly use all 4 cores and all 8 threads of the i7-920 will show that gap close. But even in those cases the i3-6100 is still just as fast as the i7-920.

However, most of what people do with computers (including playing games) only uses 1 or 2 cores anyway, making the i3-6100 far and away faster.

It also does it with a TDP of 51 watts vs TDP of 130 watts for the i7. It also comes with the benefit of a much newer motherboard that has modern features from newer/faster PCI-E slots to better DMI speeds and better USB support.

I needed more cores for rendering and greater memory support.

For that type of workload, clearly an i3 would be a silly choice. But that isn't what most people do. :)

Yes, the Haswell-E CPU is the right choice for that. For gaming, the i3-6100 is actually a shockingly good choice. If concerns over future quad core needs are there, the i5-6500 would be the next choice above that one.

8 years after this chip is released and the current chip isn't even twice as fast!! Silicon CPU progress has almost ground to a halt :-(

http://www.cpu-world.com/Compa...

The Intel Core i5 6500 (no HyperThreading) appears to be faster than AMD's FX-6300 in single-process benchmarks by 40-100%, yet the multi-process benchmarks don't show as much of a pronounced difference: 10-40%. Granted the scores/graphs are all unmarked, which limits their usefulness in trying to make an informed decision, when you don't know the actual time-frame, nor if the FX-6300 was overclocked (which it can be, and the i5 can't). As you stated, TFA compares it to a similarly priced AMD processor, as opposed to the FX-6300.

I upgraded to an AMD FX-6300 6-Core Processor Black Edition ($84.99) last year to replace a Phenom II X2 560 ($99.99, from 2011). I doubt it was actually necessary, but the FX-6300 was (and is still?) the sweet spot for price/performance.

PCIe is built in the CPU, even in $40 consumer ones. I guess that's not too bad. Else have a look at next-gen Intel high-end, that goes full-on "Itanium is dead".
One variant has what was described as a "100G" interconnect, so that's something you might use though that's yet another interconnect.

http://www.cpu-world.com/news_...

The Skylake-F will incorporate one link of first generation Storm Lake Fabric. The Skylake-F will use Multi-Chip packaging, that will house two dies, a Skylake die and a Storm Lake die. In all other aspects, the Skylake-F will be identical to the Skylake-EP.

The chips themselves (E3845, a very low power-consuming model) only consume 10W each, and it's 88 cores (not processors), each processor having 4 cores, so 22 processors consuming 220W. Assumedly these boards are designed for minimal power consumption and provide only a bare minimum of capabilities. The hard drives are SSD, so minimal power consumption there. There doesn't appear to be any sort of graphics, so no power consumption needed for that. RAM isn't usually a big power consumer. So I don't see any reason to think that the system is going to be consuming vast amounts of power. Maybe about the same total as a high-end gaming PC (but with a *lot* more computing power).

On the other hand, with epoxy-encased hardware, I am quite curious as to how they're handling cooling. I'm guessing that while the boards are embedded in epoxy that the surfaces of the processors aren't? And I'd wager that the processors are linked up to a common heat sink, as it'd make no sense to give each one its own fan. Hmm, if they're all connected to the same big chunk of aluminum or copper running all the way through, then it could double as a structural element. Neat. :)

Beyond all of this, the system is battery powered, so we're not looking at any "surge" at all. It's designed for 8 hours of usage at full power or a week of standby. The difference between the two says to me that they do a lot to power down hardware when it's not in use. Also, that battery alone probably weighs about 10-20kg** - although size-wise it's probably only 5-10 liters**, so it should fit fine.

You know, the more I look at these numbers, the more I think it all pans out. You're talking a big, heavy suitcase, but nonetheless, a suitcase.

** - assuming 8h@250-500W@200Wh/kg@400Wh/L or comparable.

AMD AM1 processors run at 25W TDP, which is less the Intel i3 processors. My current AMD AM3+ quad core processor runs at 95W TDP. I'm thinking about replacing the motherboard and getting an AMD eight-core processor that also runs at 95W TDP.

The laptops are based on the Celeron N2840, with 2GB of RAM. I can't seem to find much in the way of benchmarks; but I suspect that they are surprisingly adequate. What is a bit surprising is that the the N2840 has a quoted tray price of $107, so either Intel is cutting HP one hell of a deal, or I don't even want to know what HP cobbled the rest of the system together from...

I don't think that tray price has much basis in reality. The "$107" N2840 looks, at least on the face, to be not vastly different from the "$86" 1037U. If Biostar can sell a motherboard + 1037U + heatsink + fan for $79.99, it doesn't take much of a stretch to think maybe these prices are just "list" prices with no basis in reality. Biostar is just selling a bare motherboard so there can't be any Microsoft kickbacks or ad revenue.

The laptops are based on the Celeron N2840, with 2GB of RAM. I can't seem to find much in the way of benchmarks; but I suspect that they are surprisingly adequate. What is a bit surprising is that the the N2840 has a quoted tray price of $107, so either Intel is cutting HP one hell of a deal, or I don't even want to know what HP cobbled the rest of the system together from...

I don't think that tray price has much basis in reality. The "$107" N2840 looks, at least on the face, to be not vastly different from the "$86" 1037U. If Biostar can sell a motherboard + 1037U + heatsink + fan for $79.99, it doesn't take much of a stretch to think maybe these prices are just "list" prices with no basis in reality. Biostar is just selling a bare motherboard so there can't be any Microsoft kickbacks or ad revenue.

That's more expensive than the total cost of many tablets
http://www.cpu-world.com/CPUs/...

Although it is indeed impressive and very competitive, the Bay Trail Atom Z3770 it bests is half a year old. One would expect one of the recently announced refresh Intel parts to perform as good as or better than the Z3770 on the CPU limited benchmarks: http://www.cpu-world.com/news_...

When looking at GPU-dependent benchmarks, Intel is definitely hurting their necks while looking up to their competition and probably will be for some time.

All in all: This is good news for the consumers.

Basically, it seems to work for a brief amount of time ... Maybe five minutes, maybe a couple days, but ultimately, X will hard lock and the screen will go black

I've watched a couple hundred hours of HD video on a sandybridge quadcore desktop running Mint 14+VLC. You and/or MythTV are just doing it wrong. The symptoms you describe (intermittent hard lockups) are indicative of excessive heat or inadequate power supply, common afflictions of "media center" hardware with tiny, low quality PSUs and next to no airflow.

The set-tops you're trying emulate are engineered by actual engineers, not forum posts. They don't use 65-75w TDB processors and pretend that the system is idle while rendering HD video. They use much lower power chips and carefully engineer thermal dissipation.

The transistors saved by sharing the FPU are returned to you in integer cores, arguably better use of the transistor budget.

Arguable only if you're comparing AMD's pre-Bulldozer core designs to AMD's Bulldozer / Piledriver core designs. Whether your metric is "performance per transistor" or "performance per mm^2 of chip area" or "performance per watt", AMD's BD/PD cores beat AMD's older stuff but lose quite badly to Intel. (Even on integer code.)

Here's a more nuanced view:

Looking at the results for the two core, four thread i3-3220, the picture is more interesting, and dare I say surprising. At one extreme, the A10 is 12% worse than the i3 in an Excel test. At the other end, it outperforms the i3 by around 42% in TrueCrypt AES Encryption. Overall, the A10 shows a 4.4% performance boost over Intel's i3-3220, a figure that is close enough to call the two equal on a diverse CPU workload. This surprised me due to the fact the i3 showed very good single-threaded performance and the A10/A8 only have 2 FP units.

That AES test result is because Intel plays dumb market segmenting games with disabling AES-NI acceleration instructions in low end processors like the i3-3220. It's not because the Piledriver architecture is that good. Unfortunately the "more nuanced view" there is all about presenting an AMD-friendly opinion rather than reporting much about benchmark methodology or a complete set of numbers. You can't even tell what the average "performance boost" would be if you removed that benchmark from the mix, or how the numbers were averaged together (there are good and bad ways of doing that), and so forth. Plenty of other benchmarks 'round the web have established that one Intel core is generally more than a match for one AMD module.

Steamroller apparently only shares the MMX/FMAC unit, while standard floating point operations are per-core.

Uh, no. Steamroller retains the shared-FPU arrangement. In fact, you effectively said as much: "FMAC" operations are standard floating point operations. It doesn't get much more fundamental than multiplication and addition.

Intelligent scheduling could optimize this arrangement nicely.

Meaningless buzzwords can frabjulate the frobozz nicely! Dude, at the end of the day the FPU throughput of a Bulldozer / Piledriver module (two "cores") is still equivalent to the FPU throughput of a single Intel core. No amount of scheduling (what kind of scheduling? Compiler instruction scheduling? Hardware? Process scheduler thread placement?) will magically turn this into an advantage.

On the whole it looks to me like I'll be AMD again for my next box. Lots of cores for little $$$, yummy.

I'm betting that like most AMD fans you have no clear idea whether this would actually benefit you in any meaningful way, you just want more cores. Because obviously more is better.

6200 series have shared FPU (floating point unit). Which means that there are less FPUs that there are processing cores. To multiply two floating point numbers cores are waiting in a queue until FPU is free to use, this happens when all cores are calculating at the same time. If you are doing intensive calculations this is going to be slower than if you used 6100 series. 6100 series have dedicated FPU for each core.

The transistors saved by sharing the FPU are returned to you in integer cores, arguably better use of the transistor budget. Here's a more nuanced view:

Looking at the results for the two core, four thread i3-3220, the picture is more interesting, and dare I say surprising. At one extreme, the A10 is 12% worse than the i3 in an Excel test. At the other end, it outperforms the i3 by around 42% in TrueCrypt AES Encryption. Overall, the A10 shows a 4.4% performance boost over Intel's i3-3220, a figure that is close enough to call the two equal on a diverse CPU workload. This surprised me due to the fact the i3 showed very good single-threaded performance and the A10/A8 only have 2 FP units.

how about 6300 series, is there a dedicated FPU?

Steamroller apparently only shares the MMX/FMAC unit, while standard floating point operations are per-core. Intelligent scheduling could optimize this arrangement nicely.

On the whole it looks to me like I'll be AMD again for my next box. Lots of cores for little $$$, yummy.

Where's the FX-8170? I can't buy it if it's not produced.

All currently sold netbook/nettop Atoms are 64-bit. Yes, you can get 32-bit embedded Atoms. See here. Click on the 32-bit filter. You'll notice all models starting with E, those are embedded ones. Ignore them. N270 and N280 are from 2008 and I don't think you can get those anymore. The ones starting with Z are, like the E series, thought for non-PC platforms.

Those starting with N are for "notebooks" (netbooks, whatever the name du jour), those starting D are for "desktops" (nettops...) 64-bit all of them (except for the above noted N270 abd N280)

I can't even look at what Stanford is trying to do right now, but there have existed for years at least two online CPU "museums" that serve this goal. The one that readily springs easiest to mind - the one I've used most myself - is CPU-World. It has extensive coverage of all the major CPU lineages, including photos submitted by users, and even includes some non-CPU silicon. It seems to be largely the creation of one guy, Gennadiy Shvets, with eager collaboration from a lot of fellow enthusiasts, and there seems to be no profit motive to the site that I've ever noticed. He even thanks the most prolific contributors by name.

WHY would Stanford feel it was necessary to "divide and conquer" this enthusiasm by creating an entirely new site and museum, rather than focusing the collective interest by contributing to or partnering with the one(s) that have already existed for many years? On the face of it this effort looks like either ignorance or pointless competition.

Actually, many people have tried and 4GB works, but you only get 3.5GB because of chipset limitations. I've got a few machines like that. Not MacBooks, but just recently I had one of those early MacBooks on my workbench and I checked the possibility of upgrade online for that particular model. Official is 2GB, non-official 4GB.

It really is hit 'n miss. For example, the dumpster-sourced machine my sister uses based on a Intel D945CGCCR. According to the manual it supports 2GB. Yet, I found an older manual where it claims 4GB. This was corrected in a later manual. I suspect, that it's because the chipset doesn't allow more than 3.5GB even in 64-bit mode. (it's an E6600, it can do 64-bit). Online forums came to the consensus that 4GB works (again, with the 3.5GB limit).

My 2003 purchased AMD Athlon 2400+ MP on a Tyan Tiger 2466MPX also does this. 4GB allowed, but only 3.5GB usable *whatever you do*. As this isn't a 64-bit machine, you'd expect PAE to give you access to more. It doesn't, it is -again- a chipset limitation.

Also, older DDR1 latops are often rated 1GB... In all those I tried, replacing with 2x1GB instead of 2x512GB, worked just fine. You cannot be certain, of course, but by now I just have test modules lying around...

There is exactly one machine where I didn't manage to get it through the limit. It's a Fujitsu-Siemens Amilo Pa 1510. It's rated 2GB (DDR2) and it really doesn't want to do more. I tried 2x2GB with it (I had those lying around from another unfinished project), and it booted up with... 2.4GB. A bit more than the usual 1.8GB (~256MB for integrated graphics). So there was a win, but not really as much as a 2GB to 3.5GB upgrade.

Now, of course, I admit I'm at an advantage because I'm a dumpster diver and have parts lying around I can test before buying anything online. Still, check what is being said about such upgrades "on the street". A good starting point is here. Good luck.

The i7-2700k will have a launch MSRP of $331.

I fully expect I'll be able to get one at Microcenter for $280 or so.

Nope... The D4nn is a single core CPU, with Hyperthreading. The dual core version is the D5nn, also with Hyperthreading, I'm typing this on one (Atom D525, 199€ nettop with 2GB RAM and 320GB HDD). It's a fine little machine, but don't expect too much of it. I'm running Ubuntu on it and Flash games are unbearable. Flash video seems to be fine though. Of course, Adobe is to blame as my Core2Duo at work has the same troubles, except the problems are fewer, or more quickly resolved. (Browser locking for several seconds. Killing the plugin makes it run again, so it definitely is the plugin) While Linux can make weaker hardware function fine, once Flash is involved, performance goes through the drain. The same machine running Windows XP with the Flash plugin would run the same games fine. I have tested this myself.

Anyway, the Atom is a fine choice for basic web surfing, and office tasks. It's also a nice choice for a home server. They are best coupled with Windows (XP) due to the Flash performance issue. If Flash on Linux were better, I'd say they make perfect "daily use" machines.

These Atoms sometimes can surprise with what they can do. I have VirtualBox running on my D525 with an XP Pro guest machine. Works just fine... (Well, okay, it's for occasional tests... nothing fancy or heavy use) I was surprised to find out that that CPU doesn't have hardware virtualization, yet it works.

Nope... The D4nn is a single core CPU, with Hyperthreading. The dual core version is the D5nn, also with Hyperthreading, I'm typing this on one (Atom D525, 199€ nettop with 2GB RAM and 320GB HDD). It's a fine little machine, but don't expect too much of it. I'm running Ubuntu on it and Flash games are unbearable. Flash video seems to be fine though. Of course, Adobe is to blame as my Core2Duo at work has the same troubles, except the problems are fewer, or more quickly resolved. (Browser locking for several seconds. Killing the plugin makes it run again, so it definitely is the plugin) While Linux can make weaker hardware function fine, once Flash is involved, performance goes through the drain. The same machine running Windows XP with the Flash plugin would run the same games fine. I have tested this myself.

Anyway, the Atom is a fine choice for basic web surfing, and office tasks. It's also a nice choice for a home server. They are best coupled with Windows (XP) due to the Flash performance issue. If Flash on Linux were better, I'd say they make perfect "daily use" machines.

These Atoms sometimes can surprise with what they can do. I have VirtualBox running on my D525 with an XP Pro guest machine. Works just fine... (Well, okay, it's for occasional tests... nothing fancy or heavy use) I was surprised to find out that that CPU doesn't have hardware virtualization, yet it works.

Intel will offer Intel-branded cooling solutions for the new chips, they just won't package them with the chips.

^----- This has been confirmed: "Intel has decided to offer own brand coolers for the platform, it's just that they won't come in the box with the CPU."

So Intel will offer coolers, they're just sold separately, probably because these are cpus designed for enthusiast ("The E range (which stands for ‘enthusiast’") so they're meant for people that overclock and buy separate coolers rather than use the "stock" cooler that comes with the cpu.

Pricing of the CPUs has also been released:
_name__core__threads__freq__turbo freq__L3__TDP__price_
Core i7-3820 4 8 3.6 GHz 3.9 GHz 10 MB 130 Watt $294
Core i7-3930K 6 12 3.2 GHz 3.8 GHz 12 MB 130 Watt $583
Core i7-3960X 6 12 3.3 GHz 3.9 GHz 15 MB 130 Watt $999

http://www.cpu-world.com/CPUs/Core_2/Intel-Core%202%20Duo%20E4400%20HH80557PG0412M%20(BX80557E4400).html

EM64T, SSE 1, 2 & 3, but no VT.

Seems like you're SOL...

But 286 was a special port, and I don't think, it was capable of the features discussed: memory protection and pre-emptive multi-tasking...

Actually, I think it was. The 286 had protected mode, which Xenix used. What Xenix didn't have on 286 was decent memory paging.

Those temperatures are really dependent on the CPU we're talking about. I own a Athlon MP workstation and the CPUs routinely are in the 80 degree celcius range. Maximum is 90 degrees celcius.... That machine was rock-stable (until one of the disks started being flakey)

For more perspective, PC keyboard controllers are similar in processing power to something like this: http://www.cpu-world.com/CPUs/8048/Intel-P8048.html

That's indeed quite true: those AMD 2400+ MP CPU's were rated 90C each. I installed a passive NVidia FX5500 in the machine and running a games was 5 minutes okay, and then the graphics card started overheating. Never had that problem with my NVidia Ti4400, but that one was actively cooled.

That machine was one heck of a space heater.

Like it, the CPU is also soldered to the motherboard.

Are you sure of that? When my last laptop broke apart (the case cracked, it was over 7 years old, I think), I decided to rip it apart and sell the parts on eBay. (I never did that, I still have the parts lying around) The CPU was a P-III 600MHz mobile and it was in a ZIF socket like all CPUs (okay, these days, the pins are on the mobo, I know). It was as easy to remove it that any desktop PC, I've ever had. It's a Socket 495, specially made for mobile CPUs.

I seriously doubt laptops have soldered-on CPUs. This simply because of the fact that the same model is offered with a multitude of CPU's. Same motherboard, different CPUs. It would be too expensive to solder them on.

I'm sure it won't be easy to replace such a CPU, but soldered on? 99% sure it is not.

Your point of customization still stands, but the main reason is that the parts are extremely expensive and it's hard to open a laptop. Not because it can't be done. (I've also seen howtos on the Internet to replace discrete laptop graphics cards. No kidding.)

They both support Trusted Computing which makes me wish there was another option out there.

I don't see MOS Technology on the list...

Here's the CPU for you!

No, I promise. It *really* can be done.

I've got a pentium mmx 233 (proof of existence http://www.cpu-world.com/CPUs/80586/TYPE-Pentium%2 0MMX.html) that actually runs XP ok with 192 megs of ram.

Of course its not blazing fast, but it is usable (with a bit of patience).

KDE 3.X on the same machine requires GOBS of patience. You have to wait for minutes for the GUI to preform basic tasks.

XP can run on a surprisingly crappy system if you turn everything off, don't run on-demand AV, don't run on demand spyware scanning, and only run an app or two at a time.

Don't forget that many XP machines came from the factory with 128 MBs of ram.

Not that XP doesn't suck, a lot. Check my posting history; I flame Windows more than most. But you can shoehorn it into older PCs; I couldn't get KDE 3.X to work at a reasonable speed no matter what I did.

That's true, however it's a dirty hack. The extra memory is only addressable in 4GB (or 2GB) segments. Therefore, if you have an application that needs more than 4GB in a contiguous chunk, you are out of luck. You could probably fake it with operating system calls and complicated wrpper functions to hide the memory address arithmetic. However, you loose a lot of performance.

It's a bit like the old days of the 8086 and 80286 where memory was in 64k segments and there were segment registers used in conjunction with index registers to calculate addresses. As you can imagine, writing programs with datastructures larger than 64k was complicated and bug-prone due to the added complexity. It also slowed the program down significantly due to the extra calculations required.

So you see, a much better, cleaner, more efficient solution to the problem is to have a flat 64-bit address space, like the AMD64 architecture.

Of course, intel tried very hard to pretend that no-one "needs" 64-bits on the desktop for many years while they tried to peddle the dreadful itanium for servers and workstations. They hoped that naieve users would believe them long enough until they could get itanium PCs out in the mass market. Then along came AMD with a much better processor...

ObTrivia: One of the earliest attempts to build a supercomputer cluster out of commodity parts used a matrix of over 1024 6502 processors in parallel. As I recall, it didn't get very far, but it was probably the earliest attempt at what would be called a Beowulf-style cluster today.

Personally, I'd like to see some cheapo microprocessor in modern CD and DVD drives. Have the decoding offloaded onto the drive itself. (The Commodore PET did this via the IEEE 488 bus, in the late 1970s, You could copy one disk to another, or print a file from disk, without ever having to use the central processor.)

CPU design is fascinating, as there have been so many potentially amazing designs (such as the T400 Transputer) whose ideas were revoltionary for the time but have slowly been adopted by mainstream manufacturers.