Slashdot Mirror

← Back to Stories (view on slashdot.org)

Overclocker Pushes Intel Core i7-7700K Past 7GHz Using Liquid Nitrogen (hothardware.com)

Posted by BeauHD on from the if-you-can't-stand-the-heat-get-out-of-the-kitchen dept.

MojoKid writes from a report via HotHardware: If you've had any doubts of Intel's upcoming Kaby Lake processor's capabilities with respect to overclocking, don't fret. It's looking like even the most dedicated overclockers are going to have a blast with this series. Someone recently got a hold of an Intel Core i7-7700K chip and decided to take it for an overclocking spin. Interestingly, the motherboard used is not one of the upcoming series designed for Kaby Lake, but the chip was instead overclocked on a Z170 motherboard from ASRock (Z170M OC Formula). That bodes well for those planning to snag a Kaby Lake CPU and would rather not have to upgrade their motherboard as well. With liquid nitrogen cooling the processor, this particular chip peaked at just over 7GHz, which helped deliver a SuperPi 32M time of 4m 20s, and a wPrime 1024M time of 1m 33s. It's encouraging to see the chip breaking this clock speed, even with extreme methods, since it's a potential relative indicator of how much headroom will be available for overclocking with more standard cooling solutions.

89 of 139 comments (clear)

  1. Most depressing thing I've read all week by phantomfive · · Score: 5, Interesting

    If I recall correctly, the first time someone got over 8 Ghz was back in ~2004, over a decade ago. I know clock speed isn't everything, but parallelism will only get you so far. I really hope before we get to 5nm chips, we can get some 20 Ghz clock speeds. The amount of work you'll be able to do on a single thread will be amazing.

    --
    "First they came for the slanderers and i said nothing."
    1. Re:Most depressing thing I've read all week by Anonymous Coward · · Score: 4, Informative

      If I recall correctly, the first time someone got over 8 Ghz was back in ~2004, over a decade ago. I know clock speed isn't everything, but parallelism will only get you so far. I really hope before we get to 5nm chips, we can get some 20 Ghz clock speeds. The amount of work you'll be able to do on a single thread will be amazing.

      The only thing inherently inefficient about parallel computing is the inefficiency created by the overhead required to keep the software consistent and coherent. The real problem with multi-core computing is very little software is written in such a way that it can run on multiple CPUs. Hell, my professors were saying that in college 15 years ago, and it's still true today.

    2. Re:Most depressing thing I've read all week by phantomfive · · Score: 5, Interesting

      The only thing inherently inefficient about parallel computing

      Not all things are parallelizable. Sometimes you must wait for part A to finish before part B can begin. The obvious example is anything requiring user input (like games). Another example is databases.....when they wan to maintain ACID, they must do some things sequentially (which is unfortunately a huge bottleneck). See also, Amdahl's law.

      --
      "First they came for the slanderers and i said nothing."
    3. Re:Most depressing thing I've read all week by phantomfive · · Score: 1

      Learn C#6 ....Nothing else scales.

      Lies lol

      --
      "First they came for the slanderers and i said nothing."
    4. Re:Most depressing thing I've read all week by phantomfive · · Score: 3, Insightful

      Not everyone is putting their crap on AWS. Some of us still like our crap local. Not everything is 'in the cloud' or even 'web based.'

      --
      "First they came for the slanderers and i said nothing."
    5. Re:Most depressing thing I've read all week by Ramze · · Score: 4, Interesting

      Intel gave up on increasing clock speeds way back when they hit 4Ghz. They hit a wall, and they're done, so I wouldn't expect them to revisit it. That's when they went to multi-core. Every computer does better with dual core over single core. Most do better with quad core than dual core. (because even if a single program isn't compiled for multi-core, different programs can be assigned different cores). With VR tech and GPUs added to the cores, multi-core is likely going to continue to be the area of development for some time. As always, expect new physics and graphics extensions as well as codecs.

      Multi-core means managing the power and speed of each core individually and allowing some to power down while ramping up one or two to keep the thermal and power envelopes within tolerances. The biggest metric for Intel is performance per Watt -- as data centers are concerned about power usage for the machines and the air conditioning systems.

      I don't think there is enough of a market for enthusiasts that want 20 Ghz clock speeds for Intel to bother even doing the research for new materials to pull that off... assuming it's even possible without extreme cooling.

    6. Re:Most depressing thing I've read all week by phantomfive · · Score: 2

      Intel gave up on increasing clock speeds way back when they hit 4Ghz

      Well yeah, they 'gave up' because the technology wouldn't allow them to increase the clock speed. Halfing the die size didn't give an automatic speed boost like it had in previous generations.

      But if the next halfing of the die size does give a speed boost, do you think Intel will reject it? Of course not. A doubling of clock speed (all else equal) gives you better performance than adding an extra core.

      --
      "First they came for the slanderers and i said nothing."
    7. Re:Most depressing thing I've read all week by Anonymous Coward · · Score: 1

      Clock speed certainly isn't everything. Modern processors can perform orders of magnitude more instructions per clock than a cpu from 2004. Its the equivalent of running that 2004 cpu at 25Ghz

    8. Re:Most depressing thing I've read all week by Ramze · · Score: 5, Insightful

      This is true, but the CPU isn't the bottleneck for your examples. For user input (especially games), the user is the bottleneck. Games largely benefit from parallelization for rendering graphics. The logic isn't the bottleneck, and the latency for the response of user input is imperceptible to the human. For most instances, the RAM and CPU are waiting on the human and already have everything loaded to respond to the human. If a human's choice requires the loading of a different zone, the game could even predict which zone would load and pre-load a zone without human input, but dump it if the input wasn't what was predicted. Still, it's the I/O for the disk that's the bottleneck, not the CPU.

      As for databases, the biggest bottleneck is the storage medium. Depending on the database and how it's divided, one can even run many tasks on the same database simultaneously so long as the tables don't interact. Ramping up the CPU speed does little to nothing if the I/O to the storage medium of the database is slow b/c the db won't unlock the region of the database for the next transaction until the last transaction is written at least to a buffer if not the final storage medium.

      For that example, the best way to improve DB processing is to add RAM, add cache, and increase the clock speed of both.... if possible, even let entire tables if not the full database to exist in RAM and only write to disk periodically as a save-state. Even DDR4 2400 RAM only operates around 1.2 Ghz, though with access on rising and falling edge, it's effectively 2.4 ghz. What is your 20 Ghz CPU going to do with 10 cycles between every read and write to RAM ? Current Intel CPUs have a 4 stage pipeline. Even with a sizable cache at a higher speed, it's going to choke on the RAM latency... especially for large sequential database transactions. RAM is already hot enough to fry eggs on, so it'll be until the next RAM replacement tech comes out before we see some real boosts there. Maybe in a year or two.

      I'm curious what exactly you'd like to run at 20 Ghz through the general purpose CPU registers that can't be done better/faster with extensions using specialized hardware. For instance, x265 HEVC video playback can really heat up a CPU to nearly 100% usage, but if it has x265 decoding hardware, the CPU barely breaks 1% playing the same video on a similar CPU architecture and speed. Seems if you have a single thread that you need to have repetitively run at very high speeds, you'd rather have a FPGA or some other hardware to accommodate whatever you're trying to do rather than a general purpose cpu.

    9. Re:Most depressing thing I've read all week by Joce640k · · Score: 2

      Learn C#6 and make your code multithreaded too. It isn't going away, and people like me will eat your lunch in the coming year. Nothing else scales.

      I don't know whether to LOL or facepalm.

      People like me have already eaten that particular lunch, many years ago.

      --
      No sig today...
    10. Re:Most depressing thing I've read all week by Anonymous Coward · · Score: 1

      Not everyone is putting their crap on AWS. Some of us still like our crap local. Not everything is 'in the cloud' or even 'web based.'

      And for those of us that don't there are 10, 16, 24, and even 32 core and beyond machines out there, and they perform amazingly.

    11. Re:Most depressing thing I've read all week by Anonymous Coward · · Score: 1

      The proper response is a chuckle and a slightly condescending pat on the head, much like you would give a 5 year old who is bragging about his tic-tac-toe skills.

    12. Re:Most depressing thing I've read all week by Anonymous Coward · · Score: 2, Informative

      Frequency is not an actual measure of CPU speed, as the average number of instructions per cycle were way smaller in 2004 (on a P4 CPU) than it is today.

      Today, processors have much smaller pipelines (14 vs Prescott's 31). Pipelines contain instructions scheduled to be executed, being in different stages of executions (eg: decoded, having their necessary input ready, having their result computed in a temporary storage but not yet committed). To run a cpu efficiently, you need to keep the pipeline full. When a branch instruction (eg: "is the value of register CX 0"?) is encountered, the cpu will predict it's outcome ("no, it's not 0") and starts loading and decoding the instructions for that path in the branch to keep the pipeline full. If CX turns out to be 0, it will have to discard the entire pipeline and start to fill up with the path for the "yes, it's 0" case. Shorter pipelines reduce this penalty, resulting in a larger number of average instructions executed per cycle.

      Branch prediction also improved greatly, so you don't need to pay the already smaller penalty that many times.

      The number of cycles the execution of a single instruction takes also shrunk. An integer division took 34 cycles on a Prescott, takes 23-26 cycles on a Skylake. For simpler instructions the relative difference is even bigger, eg a MOV was at least 2 cycles for a Prescott, now it is usually 1 cycle.

      The average amount of data processed by an instruction also improved with new instructions (eg AVX, AES-NI).

      The cumulated result of these improvements (more data per instruction, more instruction per cycle, less cycles wasted in pipelines) are many times faster CPUs at the same clock rate, even without considering multiple cores.

    13. Re:Most depressing thing I've read all week by dgatwood · · Score: 1

      Uh, no. Actually, it is quite the opposite; multitrack audio processing is highly parallel.

      Let's say you have a modest 40-track project. Each project can get data from the disk independently and put it into buffers, minimizing glitches caused by one track getting delayed by another. Then each track gets all of its effects added sequentially, but each track potentially gets processed on a separate thread. Finally, the outputs of those threads get summed and additional effects added to the result, which can happen on yet another thread.

      So potentially you have 41 threads working on data in parallel, plus other threads for unrelated stuff like UI updates.

      --

      Check out my sci-fi/humor trilogy at PatriotsBooks.

    14. Re:Most depressing thing I've read all week by known_coward_69 · · Score: 1

      lots of other parts on the CPU that will give you a bigger speed boost before adding speed and cores

    15. Re:Most depressing thing I've read all week by Opportunist · · Score: 2

      Ever heard the phrase "size doesn't matter, it's knowing how to use it"? Same for raw CPU speed.

      First of all, branch prediction and pipelining has become way better in those past 10 years. Pipelines are much shorter today, and coupled with near perfect branch prediction, this alone speeds up the CPU by a factor 2 to 3. The reason for this is simply that a branch (a conditional jump, to be exact) used to mean that the CPU had to dump everything it had in its instruction pipeline and start anew from where the jump lead to. Worst case that were 30-40 cycles wasted. Per jump. I think I needn't explain how this can slow a CPU down, e.g. in tight loops where you have a handful of instructions that should (and now do) take a handful of cycles but took like 10 times as long to complete.

      That is, by the way, one of the reasons why encryption algorithms are so much faster on contemporary CPUs than they were on older models and why key sizes have quadrupled at least since 2005 to still be considered "secure".

      And of course, as many have pointed out already, the CPU isn't really the bottleneck in the current PCs. Applications we use today of course need calculations, but they are heavily dependent on periphery, if nothing else, RAM. And that is already dead slow compared to CPUs. Let's not even think about mass storage where access speeds are still measured in milliseconds. Ponder for a moment: You have a CPU running at 3 GHz, at 3 BILLION instructions per second. And you have a hard drive with an access time of 3ms (which is ... let's say faster than anything I've seen in a HDD). The CPU would have to wait for over 9000 cycles before the HD could even start to answer its request. We haven't even read anything yet, we have just waited for the HDD to spin to the point where we could start thinking about reading something from it.

      So no, the CPU isn't really the bottleneck in a contemporary PC. If you want to speed it up, get the rest of the box in gear.

      --
      We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
    16. Re:Most depressing thing I've read all week by Cassini2 · · Score: 1
      From your example, the hard drive is over 9,000,000 times slower. It's a big reason why virtual memory doesn't really work right anymore. Any quantity of misses, and a mechanical hard drive will not be able to handle the load. Even an SSD is dead slow in comparison to main RAM.

      You have a CPU running at 3 GHz, at 3 BILLION instructions per second. And you have a hard drive with an access time of 3ms (which is ... let's say faster than anything I've seen in a HDD). The CPU would have to wait for over 9,000,000 cycles before the HD could even start to answer its request.

    17. Re:Most depressing thing I've read all week by 110010001000 · · Score: 1

      It is depressing. Moores Law is dead in the CPU space. People don't realize how significant that is. Digital computing has hit a real physical dead end.

    18. Re:Most depressing thing I've read all week by Sarten-X · · Score: 1

      There will be no more automatic speed boosts.

      The reason there were jumps before as the die size decreased was that the semiconductors took less time to change states, so they could switch faster. Now, though, the parts are getting close enough together that quantum tunneling becomes a major problem, and increases in switching speed are accompanies by a drastic increase in error rates. We can still make the features smaller, but increases in speed will likely be counteracted by increases in required stabilization time.

      Rather, the biggest areas for future work are in parallelism and prediction. At the cost of complexity and chip area, we will see more predictive computations, optimizing the slowest operations into little more than lookup tables. We can also expect to see more bus channels added to chips, which in turn will give far greater I/O capability. Similarly, expect to see more features built into chips, moving toward desktops being essentially a SoC with a breakout board.

      In short, we've reached the point where raw clock rate increases are extremely difficult for only marginal gains. Increases in speed, however, are sill promising, but they come from making better use of the cycles we have.

      --
      You do not have a moral or legal right to do absolutely anything you want.
    19. Re:Most depressing thing I've read all week by JBMcB · · Score: 1

      Then each track gets all of its effects added sequentially, but each track potentially gets processed on a separate thread. Finally, the outputs of those threads get summed and additional effects added to the result, which can happen on yet another thread.

      So potentially you have 41 threads working on data in parallel, plus other threads for unrelated stuff like UI updates.

      That does speed things up, however, processing a stack of effects on a single track can take quite a long time, especially if you're doing everything in 96KHz/192KHz at 24-bit, which is how a lot of mixing is done these days before dithering down to 44.1/16.

      Also, with 41 tracks of HD audio, RAM throughput is going to be a bottleneck - so performance isn't going to scale linearly across threads. The bottleneck is the effects stack itself, which the OP pointed out, isn't very parallelize-able, so brute-force crunching with a ton of clock cycles is the best way to speed this up.

      I know a few audio engineers, and they are the only non-gamers I'm aware of who overclock their machines. They use those overbuilt, sound-deadened chassis with water cooled everything.

      --
      My Other Computer Is A Data General Nova III.
    20. Re:Most depressing thing I've read all week by Kjella · · Score: 2

      I'm curious what exactly you'd like to run at 20 Ghz through the general purpose CPU registers that can't be done better/faster with extensions using specialized hardware.

      Nothing, obviously. I'll just submit my DFS4ME (do funky shit for me) instruction to Intel and I'm sure they'll put it in the next stepping or create a special batch just for me. I can even pay $50 extra, though I need it next week. I'll also reverse engineer and patch that proprietary binary I got to use the new instruction, that totally won't be any work or void any support. Or I could buy that 20 GHz machine and have everything magically work much, much faster. Nah, I'll just do the first one.

      --
      Live today, because you never know what tomorrow brings
    21. Re: Most depressing thing I've read all week by Khyber · · Score: 1

      "I think it's extremely naive to believe audio can be processed efficiently in parallel"

      Guess you've never played with an EMU 10K1 chip before, even though those were out in the late 90s.

      --
      Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
    22. Re:Most depressing thing I've read all week by Hylandr · · Score: 1

      Assembly, C, C++

      You were saying?

      --
      ~ People that think they are better than anyone else for any reason are the cause of all the strife in the world.
    23. Re:Most depressing thing I've read all week by Hylandr · · Score: 1

      I wish I hadn't commented on this thread yet, as this needs a +1.

      You kids may not like it, but he's dead on.

      --
      ~ People that think they are better than anyone else for any reason are the cause of all the strife in the world.
    24. Re: Most depressing thing I've read all week by ranton · · Score: 1

      AWS is wonderful for R&D usage because it makes procuring money for R&D equipment far easier. My guess is you weren't part of the group which obtained the $5 million investment in your 1000 node cluster (or hypothetical cluster if your example doesn't exist). If you were then you wouldn't think it was easier than obtaining a monthly R&D budget for AWS on demand spending. Most companies prefer discretionary spending which can be adjusted later as opposed to huge up front expenditures which limit flexibility in the future.

      Of course the R&D group would prefer to not be accountable to anyone but that is rare in the real world.

      --
      -- All that is necessary for the triumph of evil is that good men do nothing. -- Edmund Burke
    25. Re: Most depressing thing I've read all week by Hylandr · · Score: 1

      This is why there's a push to develop locally with VirtualBox or your favorite VM Host whilst using the same recipes you would normally deploy to your machine provider.

      I mention VirtualBox as it's the one I am currently running despite being owned by the Ebil Oracle. If there's others you prefer besides VMWare *NOT* windows based I would love to see a list. I specified not windows because I have no intention of paying for a licence for the host hardware.

      --
      ~ People that think they are better than anyone else for any reason are the cause of all the strife in the world.
    26. Re: Most depressing thing I've read all week by griffincalme · · Score: 1

      Personally, I'm not worried at all about the stagnation of clock speeds. GHz aren't everything, instructions per clock have greatly increased even if clock speed hasn't. This means that a 4GHz Skylake single thread is much faster on the same task when compared to a 4GHz Sandy bridge single thread. See chart: https://www.nextplatform.com/w...

    27. Re:Most depressing thing I've read all week by Lord+Apathy · · Score: 1

      I don't think there is enough of a market for enthusiasts that want 20 Ghz clock speeds for Intel to bother even doing the research for new materials to pull that off... assuming it's even possible without extreme cooling.

      I don't believe this will be possible even with extreme cooling. I read a article once, might have even been from /., where the switching speeds inside the cpu where approaching the light speed barrier. Meaning the switches where changing as fast as they possibly ever could.

      I don't remember in the paper if this was a physical problem today or a theoretical problem down the road. Whatever it was, the light speed and switching speed will be the final limit on the mhz race.

      --

      Supporting World Peace Through Nuclear Pacification

    28. Re: Most depressing thing I've read all week by Midnight+Thunder · · Score: 1

      I am guessing most developers aren't that good at multi-threading. For most jobs it isn't necessary, and when it is someone will do a good enough job or find the developer or library that will take care of it.

      If you ever do multi-core GPGPU you see there is a threshold below which just using the CPU is good enough and may be faster. Part of that is that you only really benefit from the GPGPU if the processing queue is kept full.

      Then there are languages such as NodeJS which are single threaded and when you need to parralize it is via separate processes, so you really need to reduce the inter-process chatter.

      --
      Jumpstart the tartan drive.
    29. Re:Most depressing thing I've read all week by unixisc · · Score: 1

      Incidentally, this is a rehash on the discussions of superscalar vs superpipelined processing that we had in RISC CPUs in the 90s - whether more MHz were better w/ more pipeline stages - like in the Alpha and MIPS III, or more pipelines were better w/ fewer stages, but running in tandem. Today's multi-core is a macrocosm of superscalar, while what this story discusses - pumping up the GHz - is essentially trying brute force on the problem.

    30. Re:Most depressing thing I've read all week by m.dillon · · Score: 2

      Paging to a hard drive doesn't really work in this day and age, the demands of the VM system (due to the commensurant increase in scale of modern machines) are well in excess of what one or two HDDs can handle.

      However, virtual memory works quite well with a SSD. Sure, the SSD isn't as fast as memory, but the scale works similarly to how cpu caches vs main memory scaling works. Its back in the ballpark so the system as a whole works quite well.

      It depends on the workload of course... browsers are particularly bad if they exceed available ram but its primarily because browsers fragment their memory space badly. Firefox, for example, with a 4GB VSZ, will keep 3GB in core no matter what due to fragmentary access, even though it might only be using 1GB worth of memory in its actual accesses.

      For example, one of our bulk builders has 128GB of ram and roughly 200GB of SSD swap configured. In order to configure enough paralllelism to keep the 48 cores fully loaded at all times a portion of the build (when it gets to the larger C++ projects) will require more than 128GB of ram and start eating into the swap space to the tune of another ~100GB or so. However, the cpus are still able to load to 100% because there is enough parallelism to absorb the relatively fewer processes blocked on page-in.

      Similarly, my little chromebook with 4G of ram has 16GB of SSD swap configured (running DragonFly of course, not running Chrome), and has no problem with responsiveness despite digging into that swap quite extensively.

      So virtual memory does, in fact, work well. And it will work well in most use cases when configured properly with a SSD as backing store. One can also go beyond the SATA SSD and throw in a NVMe SSD for swap, which is even faster (~3GBytes/sec reading for a cheap one). Given that main memory typically has 25-50 GByte/sec of bandwidth, that's only a 8x to 16x difference in speed.

      -Matt

    31. Re:Most depressing thing I've read all week by m.dillon · · Score: 2

      The main bottleneck for a modern cpu are main memory accesses. What is amazing is that all of the prediction and the huge (192+) number of uOPS that can be on the deck at once is able to absorb enough of the massive latencies main memory accesses cause to bring the actual average IPC back towards roughly ~1.4. And this is with the cache misses causing *only* around ~6 GBytes/sec worth of main memory accesses per socket (with a maximum main memory bandwidth of around 50 GBytes/sec per socket, if I remember right).

      Without all of that stuff, a single cache miss can impose several hundred clock cycles of latency and destroy average IPC throughput.

      So, for example, here is a 16 core / 32 thread - dual socket E5-2620v4 @ 2.1 GHz system doing a bunch of parallel compiles, using Intel's PCM infrastructure to measure what the cpu threads are actually doing:

      http://apollo.backplane.com/DF...

      Remember, 32 hyperthreads here so two hyperthreads per core. Actual physical core IPC (shown at the bottom) is roughly 1.39. At 2.1-2.4 GHz this system is retiring a total of 55 billion instructions per second.

      In this particular case, being mostly integer math, the bottleneck is almost entirely memory-related. It doesn't take much to stall out a core. If I were running FP-intensive programs instead it would more likely be bottlenecked in the FP unit and not so much on main memory. Also note the temperature... barely ~40C with a standard copper heatsink and fan. Different workloads will cause different levels of cpu and memory loading.

      -Matt

    32. Re:Most depressing thing I've read all week by Lord+Apathy · · Score: 1

      It is depressing. Moores Law is dead in the CPU space. People don't realize how significant that is. Digital computing has hit a real physical dead end

      Not really. How many people really need faster or more powerful computers? Well at least for our desktops. Even that being said most servers that I have been working with have 20 ore more core but still top out between 2 and 3 ghz.

      Most computing aspects of the public can be done by a modest 2 core processor at 2.6 ghz or slower even. It doesn't take a i7 to read email, surf the web, or balance a check book. I have a 3.6 ghz i7 at work and all I use it for a glorified terminal server and place to read my email.

      My personal gaming rig is 4ghz i7 that I just built. It's a quad core. It replaces another quad core that run at 3.7 ghz. Instead of getting a faster processor with more cores I went with faster I/O in the form of a m.2 pci module. I didn't see where more cores would improve my FPS where faster I/O would make things load faster. So far I have been proven correct.

      --

      Supporting World Peace Through Nuclear Pacification

    33. Re:Most depressing thing I've read all week by fahrbot-bot · · Score: 1

      It's a big reason why virtual memory doesn't really work right anymore.

      Virtual memory wasn't/isn't to provide increased (or even sustained) speed, but rather increased overall system capacity. So it still works fine as intended.

      --
      It must have been something you assimilated. . . .
    34. Re:Most depressing thing I've read all week by parkinglot777 · · Score: 1

      The human mind is a MASSIVELY parallel computer, ...

      No, human brain is NOT parallel computing. You may think that you can do multi-tasking, but your brain is actually switching tasks for you.

      ..., but its slow as shit compared to even the first intel CPUs when it comes to complex math

      Even though it is not as fast as computers, apparently it is fast enough to make you think you are performing parallel computing. ;)

    35. Re:Most depressing thing I've read all week by lgw · · Score: 1

      This gets into semantics on "parallel". The CPU-intensive part isn't efficiently parallelized, but that rarely matters, as you spend most of your time I/O bound, and that parallelizes well.

      Most applications are this way - more than 2 cores doesn't help very much, because you're not CPU-bound in the first place. There are noteworthy exceptions, like video transcoding, where CPU is entirely the limiting factor, but those exceptions tend to parallelize well, because people have put in the effort.

      The main area where more cores could help, but the software isn't written for it, is games, because of the legacy of single-threaded game engines. This is slowly shifting, but it take a generation to completely change the way an industry thinks about a problem, and we're only half-way through that shift IMO.

      --
      Socialism: a lie told by totalitarians and believed by fools.
    36. Re:Most depressing thing I've read all week by mrchaotica · · Score: 1

      Given that the 20 GHz machine isn't actually possible with current technology, implementing your DSF4ME instruction in an ASIC really is your only option if you need it to be that fast.

      --

      "[Regarding the 'cloud,'] ownership was what made America different than Russia." -- Woz

    37. Re: Most depressing thing I've read all week by lgw · · Score: 1

      Most companies don't realize that AWS is amazing for "factory work" (stuff that's already done, you're just running on a grand scale for revenue generation). But it's terrible for R&D usage---especially the metered usage variety.

      My team recently ran some exhaustive testing to be sure an algorithm would behave as expected - 30000 core-hours. We didn't have any servers just lying idle, but it was trivial (and cheap) to run the test overnight using servers from EC2 Spot. Much better than freeing a few high-end servers for a week.

      Really, any sort of temporary usage, cloud is king. No pre-provisioning, six-week procurement delays, or budget arguments that "no really, we'll use these machines for a lot of different problems to come". This is why tech start-ups have shifted to doing everything that can in the cloud - nightly build and QA runs, test servers for whatever you're building. The nice upside for a start up is that if you build in in the could, even though you only have 3 customers, it's easy to tell a potential acquirer "sure it will scale to 3 million customers, cloud cloud cloud", and it might even be true.

      If you haven't caught on to the theme yet: you don't keep those metered cloud serves just sitting around doing nothing - you automate launching everything as-needed, and only pay for servers when you have work for those servers to do.

      --
      Socialism: a lie told by totalitarians and believed by fools.
    38. Re: Most depressing thing I've read all week by phantomfive · · Score: 1

      So like, 10% improvement per generation?

      --
      "First they came for the slanderers and i said nothing."
    39. Re:Most depressing thing I've read all week by phantomfive · · Score: 1

      How many people really need faster or more powerful computers?

      *raise my hand*.

      Imagine having an embedded CPU running at 10 GhZ.

      --
      "First they came for the slanderers and i said nothing."
    40. Re:Most depressing thing I've read all week by phantomfive · · Score: 1

      I'm curious what exactly you'd like to run at 20 Ghz through the general purpose CPU registers that can't be done better/faster with extensions using specialized hardware

      Anything (pretty much) can be done faster with specialized hardware (hardware RAIDs are usually slower than software RAIDs for some reason), but the more processing you can get done between frames at 50fps the better.

      --
      "First they came for the slanderers and i said nothing."
    41. Re:Most depressing thing I've read all week by swillden · · Score: 1

      A doubling of clock speed (all else equal) gives you better performance than adding an extra core.

      But does it give you better performance per watt? That's the metric they care about, because it's the metric their high-power computing customers care about. And doubling clock speed, all else equal, does not give you more performance per watt than adding an extra core.

      Take a look at really high-end CPUs (say, $2000+) and you'll see that almost none of them reach 3 GHz, and many are below 2.5 GHz. But they have a dozen-plus cores, and are designed to be used in two or four-processor motherboards. That's where the high-performance market is, because that's where you get the best performance per watt.

      --
      Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
    42. Re:Most depressing thing I've read all week by Cute+Fuzzy+Bunny · · Score: 1

      They hardly gave up or hit a wall. Hell the Pentium 4 was slated to go to 12GHz and beyond. You'd have needed the liquid nitrogen for that, or use it to heat your house.

    43. Re:Most depressing thing I've read all week by Cute+Fuzzy+Bunny · · Score: 1

      Its far from dead, and there are plenty of new materials and processes kicking around to keep it going indefinitely.

    44. Re:Most depressing thing I've read all week by ChrisMaple · · Score: 1

      Huge numbers of processors and cores only give the highest performance if the software is designed to make effective use of those resources. Typical amateur software and most freeware is single-threaded, and that software benefits from high clock rates and high instructions-per-clock.

      The design choices that allow large numbers of moderate-power cores differ from the choices that would allow one or two extremely fast cores, and Intel isn't about to design a different core that would optimize single thread speed. For home users that aren't running multi-threaded games, this direction of development is not in our favor.

      --
      Contribute to civilization: ari.aynrand.org/donate
    45. Re:Most depressing thing I've read all week by ChrisMaple · · Score: 1

      Not the switching speed, but the transit time from one section of the chip to another. At 20 GHz, a signal will go less than 1 cm per cycle.

      --
      Contribute to civilization: ari.aynrand.org/donate
    46. Re:Most depressing thing I've read all week by ChrisMaple · · Score: 2

      Intel is sill cheaping out on the thermal interface goo. People who want top performance are having to de-lid the CPUs and use good stuff.

      In order to save 10 cents on a $200+ CPU, Intel is deeply cutting potential performance and increasing the thermal stress on the chips. More thermal stress means shorter time to failure.

      --
      Contribute to civilization: ari.aynrand.org/donate
    47. Re:Most depressing thing I've read all week by ChrisMaple · · Score: 1

      Orders of magnitude? Over 100 times more IPC? Not bloody likely.

      --
      Contribute to civilization: ari.aynrand.org/donate
    48. Re:Most depressing thing I've read all week by Cederic · · Score: 1

      Fucking hell. You deal in small small databases.

      When you're looking for terabytes of RAM for your database it's not the fucking SQL that's the issue, trust me. Shit, there's a good chance your database doesn't even support SQL.

    49. Re: Most depressing thing I've read all week by Cederic · · Score: 1

      Can confirm what the other three replies said. R&D that needs one-off workloads is an ideal AWS use case.

      Others include swift onboarding of new workloads, where in-house can't move quickly enough.

      Another is the processing of customer data that they already have on AWS - it's cheaper to pre-process it there and only pull down the results.

      But yeah, you could also put your factory work on there. Between them Azure and AWS are seriously increasing the size of business needed before owning a data centre remotely makes sense.

    50. Re:Most depressing thing I've read all week by Cederic · · Score: 1

      FM17 still hits 100% CPU on one core for well over a minute during new game initiation. Yeah, I'd like that to go faster. I already have a m.2 pci drive, I have lovely fast RAM and I have a CPU running at 4.2GHz. It's not fast enough.

    51. Re:Most depressing thing I've read all week by swillden · · Score: 1

      Huge numbers of processors and cores only give the highest performance if the software is designed to make effective use of those resources. Typical amateur software and most freeware is single-threaded, and that software benefits from high clock rates and high instructions-per-clock.

      The design choices that allow large numbers of moderate-power cores differ from the choices that would allow one or two extremely fast cores, and Intel isn't about to design a different core that would optimize single thread speed. For home users that aren't running multi-threaded games, this direction of development is not in our favor.

      For home users, computers got fast enough years ago. Consumers are pushing for thinner and lighter, longer battery life, etc., not more horsepower. The customers demanding power are the big data centers... and they do know how to make effective use of large numbers of cores, and their focus -- as I said previously -- is on performance per watt, not raw performance.

      Heh. I should mention that the desktop in my home office has 40 cores @ 2.6 Ghz. My heavy-duty processing requirements are for compiling large software packages, and that is a highly parallelizable problem (I typically do make -j80). The other bottleneck when building software is disk I/O, which is why I have 1.25 TB of SSD.

      --
      Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
    52. Re:Most depressing thing I've read all week by dgatwood · · Score: 1

      Your knowledge of computer sound processing appears to be quite limited and rather obsolete. Although you are correct that there is a lot of highly parallel processing in high performance multitracking, as others point out each individual thread still has the highest performance demands imaginable. Some software synthesizers use state of the art circuit simulation software and increase the fidelity as much as they can given the clock cycles they have to work with. Get a faster cpu and you will immediately and noticeably improve the quality of your output in many situations. The faster we can go, the more synthesis techniques we can use that have not even been dreamt of today because we don't have the power to use them in realtime.

      If you're getting lower quality mixes because of a lack of CPU power, you're doing it wrong. Yes, some software has a quality setting to let you reduce CPU overhead, but there's no reason to ever use anything but the highest setting during final mixdown, because that is not real-time.

      --

      Check out my sci-fi/humor trilogy at PatriotsBooks.

    53. Re:Most depressing thing I've read all week by dgatwood · · Score: 1

      The CPU-intensive part isn't efficiently parallelized, but that rarely matters, as you spend most of your time I/O bound, and that parallelizes well.

      Actually, even if you're only dealing with post-mix effects, in many cases, the CPU-intensive part is also efficiently parallelized. FFTs, for example, are highly parallelizable.

      --

      Check out my sci-fi/humor trilogy at PatriotsBooks.

    54. Re:Most depressing thing I've read all week by dgatwood · · Score: 1

      Also, with 41 tracks of HD audio, RAM throughput is going to be a bottleneck - so performance isn't going to scale linearly across threads. The bottleneck is the effects stack itself, which the OP pointed out, isn't very parallelize-able, so brute-force crunching with a ton of clock cycles is the best way to speed this up.

      My quad G5 did a decent enough job with 40 tracks of 96/24 audio, and that was over a decade ago. Forty tracks shouldn't even be considered heavy lifting for any modern four-core CPU. (192 kHz audio is just wasting disk space, unless being able to decode WWVB in post-production is a virtue for some reason....) Besides, most of the more complex effects plug-ins take advantage of libraries like Accelerate framework or BLAS to parallelize a lot of the processing, too.

      --

      Check out my sci-fi/humor trilogy at PatriotsBooks.

    55. Re: Most depressing thing I've read all week by Khyber · · Score: 1

      Perhaps you should, oh, I dunno, own one, get Linux, and get the kX driver set. And them maybe you'll understand that all of the effects it does are processed in parallel, real-time.

      --
      Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
    56. Re: Most depressing thing I've read all week by Khyber · · Score: 1

      Nope, your dumb ass is owned, thanks to /. handing out IP addresses of posts when prompted to by a court. See? You're so stupid that you can't even be bothered to identify yourself, but I already know who you are, and so do my friends.

      --
      Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
  2. Back in Reality by negRo_slim · · Score: 1

    Some minor gains but an otherwise mediocre cycle.

    Still I'm anxious to upgrade my Core i7-3630QM. This looks good enough.

    --
    On the Oregon Cost born and raised, On the beach is where I spent most of my days
    1. Re:Back in Reality by wbr1 · · Score: 2

      Wait for Zen. Should be good. Intel's repeated 5% gains are not enough to drive an upgrade cycle for me or anyone I recommend to. 3-4Gen is still more than adequate for most tasks.

      --
      Silence is a state of mime.
    2. Re:Back in Reality by 0100010001010011 · · Score: 1

      If you're looking for a small bump in the mean time you can get a i7-3940XM. It shares the rPGA988B socket with the 3630QM. They're ~$300 on eBay.

      In Benchmarks it's ~20% faster than what you have currently.

      It's still faster than the 6820HQ (8765/1891) or E3-1505M (8699/1887) CPUs that Dell ship in their newest 7000 Precision lines.

  3. My be scared of the Zen of whoopass by future+assassin · · Score: 1

    coming soon.

    --
    by TheSpoom (715771) Uncaring Linux user here. I have nothing to add to this but please continue. *munches popcorn*
    1. Re:My be scared of the Zen of whoopass by Lord+Apathy · · Score: 1

      I believe that I speak for a lot of us when I say we'll believe it when we see it. The specs for Zen are look impressive when compared to the old Bulldozer cores. But compared to some of the current Intel cores, not so much

      I remember the specs that first came out for the Bulldozer cores. They looked pretty impressive at the time. But in real life they didn't live up to those specs. An this I know. I had two AMD bulldozer systems at the same time.

      I do hope you are correct in your enthusiastic comment. I want to see a strong AMD to offset Intel but I will hold off till I see some Real World benchmarks on Zen.

      --

      Supporting World Peace Through Nuclear Pacification

    2. Re:My be scared of the Zen of whoopass by Cute+Fuzzy+Bunny · · Score: 1

      There isn't any such thing as a 'strong amd to offset intel'.

      Intel has an extremely efficient in house manufacturing division and billions of cash in the bank. At any given moment they can drop their prices, throw in some new features and submarine AMD, Having to farm out manufacturing to outfits that are nowhere near as efficient and losing money for over a decade aren't a good platform.

      The thing is, AMD has never really meant much to Intel. Since the Pentium 4 days, the only thing AMD did was push intel to release 64 bit chips. Which then sat without user models or software to take advantage of it. But hey, users love more bits, whether they're useful or not.

      They're a nice foil to fend off anti trust suits. Other than that...

      I didn't think much of AMD recently but when they released the 480 and overdrew the power to the board past PCIE specs to get a little more oomph out of a disappointing overhyped release, I really sensed the desperation. Overpromise, under-deliver and disappoint. I expect Zen will be a fine product that uses more power, makes more heat, and performs below its Intel counterparts.

    3. Re:My be scared of the Zen of whoopass by Lord+Apathy · · Score: 1

      I expect Zen will be a fine product that uses more power, makes more heat, and performs below its Intel counterparts.

      I believe you just said what all of us are thinking.

      --

      Supporting World Peace Through Nuclear Pacification

  4. umpf... by 4im · · Score: 2

    I googled for "highest cpu clock speed" and got e.g. http://valid.x86.fr/records.html

    It seems this is a far cry from what's been done elsewhere, with numbers there showing over 8.5GHz.

    Anyway, my criteria are rather low-energy, low-noise computers than extreme clock frequencies, even if I can make use of them.

  5. Meanwhile, the bus speed... by inflex · · Score: 4, Informative

    ... lumbers along at 100MHz still.

    Time we started working on that side of the hardware some more.

    1. Re:Meanwhile, the bus speed... by drolli · · Score: 1

      well, then we need another kind of connector with much tighter tolerances/mroe expensive manufacturing processes

    2. Re:Meanwhile, the bus speed... by fahrbot-bot · · Score: 1

      well, then we need another kind of connector with much tighter tolerances/mroe expensive manufacturing processes

      Cray (reportedly) did this with their systems. For example, they would measure individual chip speed when building their memory boards and put faster chips farther from the connector and slower chips closer. Of, course the Cray 2 cost about $17 million back in the day -- I was one of the admins on voyager at NASA LaRC in the late 1980s. '88 Cray Too Old For Nasa

      --
      It must have been something you assimilated. . . .
    3. Re:Meanwhile, the bus speed... by thegarbz · · Score: 1

      ... lumbers along at 100MHz still.

      Time we started working on that side of the hardware some more.

      Huh? Bus speeds haven't been that slow since the 90s, and they aren't the limiting factor in CPU intensive operations.

  6. Slightly offtopic by Artem+S.+Tashkinov · · Score: 2
    Let me copy this comment since I believe it's kinda relevant.

    Kaby Lake is an embarrassment and should have never seen the light of day - at least not as a "new" architecture - it's anything but "new". The CPU core is completely the same, the GPU core is minimally improved. Intel should have released it as SkyLake with XX50 postfix (6750/6650/etc) because it's what it is.

    I'm quite sure it's Intel's marketing department and investors who insisted that Intel should release something "new". I want to believe Ryzen will pan out to give a good blow to Intel and then we'll have some real improvement/competition.

    1. Re:Slightly offtopic by phantomfive · · Score: 1

      What kind of 'new' stuff do you want? I don't really want anything new from Intel's architecture, I just want a faster chip. New instructions seem dumb to me (but I'm sure some people want them).

      --
      "First they came for the slanderers and i said nothing."
    2. Re:Slightly offtopic by known_coward_69 · · Score: 1

      this is how science and engineering work stupid

      small improvements and over time it's a big leap. like Apple's CPU's. 6 years ago they were just another reference design. little changes every year and now they are faster than some of the desktop CPU's on the market

    3. Re:Slightly offtopic by 0100010001010011 · · Score: 1

      I just want a faster chip

      New instructions seem dumb to me

      That's exactly how they get faster chips with the 'same' clock speed.

      Some embedded chips can only do int16 adds/multiplies in hardware and rely on software tricks to do floating point math, they're slow compared to a chip than can do the floating points adds/multiplies in hardware.

      Compare LINPACK benchmarks using AVX2, AVX and SSE.

    4. Re:Slightly offtopic by TeknoHog · · Score: 1

      What kind of 'new' stuff do you want? I don't really want anything new from Intel's architecture, I just want a faster chip. New instructions seem dumb to me (but I'm sure some people want them).

      I just want a faster horse.

      Sincerely Yours,
      Henry Ford

      --
      Escher was the first MC and Giger invented the HR department.
    5. Re:Slightly offtopic by Cute+Fuzzy+Bunny · · Score: 1

      I want a pony!

  7. Computers are still too slow by Bruinwar · · Score: 2

    My dream (fantasy?!) is for computers (& networks!!) that are faster than me. It should be waiting on me, not me waiting on it. All day long I wait on my computer. They are not even close to being fast enough. It doesn't help at all that the software seems to be slower than ever.

    --
    SLOWER TRAFFIC KEEP RIGHT
    1. Re:Computers are still too slow by yuvcifjt · · Score: 2

      Erm, pretty stupid comment if you just think for a few seconds...
      Unless you're encoding video or 3d-rendering all-day-long, for most people, the CPU is almost always idle for everyone else and just wasting power waiting for human input.

      Even when compiling code, or rendering a web page, or doing a database transaction, the CPU wakes a couple of cores for a few seconds before it returns to almost idle.

      And by the way, if you see the CPU meter at 100%, it likely that it's not the CPU that's the bottleneck, but rather the memory (or the bus), especially when gaming.

    2. Re:Computers are still too slow by Bruinwar · · Score: 1

      I've gave it more than a few seconds of thought, pretty much on a daily basis while sitting there waiting. This may be slightly off topic because the article is about clock speeds, but what I said, if you read carefully, is I am waiting on a computer (& network). Of course it's a lot of things combined that causes me to be waiting.

      My personal desktop runs a i7 at 4GHz, 16 gig RAM, two 500 gig SSDs striped, an a nVidia 960. Runs like a champ, but when using it for work, I still find myself waiting.

      A lot of things need to be improved to get where I would like computers to be. Higher clock speeds would be one of them.

      --
      SLOWER TRAFFIC KEEP RIGHT
  8. As much as I like overclocking... by Chas · · Score: 1

    I look at LN2 overclocking, shutting down all but a core or two, etc, as basically masturbation using lidocaine jelly as lube.

    Sure, it produces an interesting result.

    But, in the end, it's ultimately useless for purpose.

    --


    Chas - The one, the only.
    THANK GOD!!!
  9. Re:Zen pci-e lanes will crush intel and no DMI lin by Holi · · Score: 1

    I hope your right, but AMD has failed to wow since the Ghz race ended.

    --
    Sorry, teleporters just kill you and then make a copy. A perfect, soul-less copy.
  10. dumb headline by slashmydots · · Score: 1

    They probably turned off all the other cores and it says they used liquid nitrogen. So yes, I am still worried about its overclocking capabilities in real world usage. Also 7GHz isn't that good.

  11. parallelism vs raw clock speed by unixisc · · Score: 1

    The real problem with multicore computing is it doesn't solve every problem. Some problems are inherently serial and no number of course or software solutions are going to change that.

    But ain't that precisely the reason for what the GP said?

    The real problem with multi-core computing is very little software is written in such a way that it can run on multiple CPUs.

    As you pointed out, a lot of problems are inherently serial, and can't be resolved by software. VLIW was one attempt in the past to try resolving it by tossing everything to a compiler, but it turned out not only that such feats were easier said than done, but also, that the actual real estate saved in CPUs turned out to be much less than anticipated.

    TFA - I really don't think people should try overclocking i7s. Instead, if their workloads need plenty of compute power and are not divisible into different cores, they should try working on interleaved workloads, where one task is handled by each core. Intel does have CPUs up to 72 cores, which one can buy if they have the need - and the budget

    1. Re:parallelism vs raw clock speed by m.dillon · · Score: 2

      The unfortunate result of VLIW was that the cpu caches became ineffective, causing long latencies or requiring a much larger cache. In otherwords, non-competitive given the same cache size. This is also one of the reasons why ARM has been having such a tough time catching up to Intel (though maybe there is light at the end of the tunnel there, finally, after years and years). Even though Intel's instruction set requires some significant decoding to convert to uOPS internally, it's actually highly compact in terms of the L2/L3 cache footprint. That turned out to matter more.

      People often misinterpret the effects of serialization. It's actually a matrix. When a portion of a problem has to be serialized it winds up adding latency, but requiring a portion of a problem to be serialized does not necessarily mean that the larger program cannot run with parallelism. There will often be many individual threads each having to run serially which in aggregate can run in parallel on a machine, and in such cases (really the vast majority of cases) one can utilize however many cores the machine has relatively efficiently.

      This is true for databases, video processing, sound processing, and many other work loads. For example, if one cannot parallelize video compression on a frame by frame basis that doesn't mean that one cannot use all available cpus by having each cpu encode a different portion of the video.

      Same with sound processing. If one is mixing 40 channels in an inherently serialized process this does not prevent the program from using all available cpus by having each one mix a different portion of the overall piece.

      For databases there will often be many clients. Even if the query from one particular client cannot be parallelized, if one has 1000 queries running on a 72-core system one gains scale from those 72 cores. And at that point it just comes down to making sure the caches are large enough (including main memory) such that all the cpus can remain fully loaded.

      -Matt

    2. Re:parallelism vs raw clock speed by Cederic · · Score: 1

      TFA - I really don't think people should try overclocking i7s.

      Why deny people their hobby? Why force people to buy a more expensive chip when they can get good compute performance from their cheaper one?

      There's a whole market out there for overclocking support, it's a mini industry in its own right. Sure, you could just buy higher compute capability by throwing money at the problem but where's the fun in that?

    3. Re:parallelism vs raw clock speed by unixisc · · Score: 1

      So does the excess cooling caused by overclocking come for free, or less than the incremental performance price that the next higher CPU will give? Not to mention the excess power consumption

  12. Re:Zen pci-e lanes will crush intel and no DMI lin by Lonewolf666 · · Score: 1

    Rumors (because that's what they are at this point) talk about 32 PCIe lanes for desktop Zen.
    That is good but not exceptional, the better Intel E-series CPUs have 40.

    But Zen will probably beat the E-series in price while coming (very) close in performance, much better than the situation with Bulldozer. That should be enough to shake up the market, please the AMD fans and maybe force Intel to rethink its very high prices.

    In short, I don't expect Zen to be a wonder CPU, but I expect it to bring real competition back.

    --
    C - the footgun of programming languages
  13. Re:Imbiciles by unixisc · · Score: 1

    Why? When they are not warranty'ing 7GHz, why should they supply parts that do nothing but add to the cost? If someone is so unique that they need that 7GHz, they need to separately buy those cooling units. And maybe move to Alaska or Nunavut while they are at it, and keep the CPU outdoors under a shed on a roof so that it will always have wind blowing in one direction or another, w/o having the snow land on the CPU.

  14. Re:Zen pci-e lanes will crush intel and no DMI lin by Lord+Apathy · · Score: 1

    I hope your right, but AMD has failed to wow since the Ghz race ended.

    I hope so. As someone that has been burned twice in the past by AMD's promises, I chose intel this time for my workstation. It will be a number of years before I will conciser switching back to AMD.

    All though I have to admit that 8350 did make a nice centos server.

    --

    Supporting World Peace Through Nuclear Pacification

  15. Re:Zen pci-e lanes will crush intel and no DMI lin by Joe_Dragon · · Score: 1

    Kaby Lake-X only has 16 + DMI

    ZEN may have 32-64 main or 32 main + upto 32 for storage / network / usb / etc. (not stacked like Intel DMI is.)