|
BangersInMyKnickers posted:Do you really need an extra 600mhz? The more voltage you crank through this thing, the shorter its lifespan will be and at some point you have to ask yourself if it is worth mucking around with any more. I understand your concern but 1.3V is by no means a high voltage for a 2500K/2600k CPU to run at. Mine is at 1.38V and it barely breaks 60C at full load. I understand there is more to this than temperature but people have even asked motherboard manufacturers and been told that voltages up to 1.42V are safe. People have stress tested these things for days with voltages over 1.5V with no issues.
|
# ¿ Mar 31, 2011 04:41 |
|
|
# ¿ Apr 23, 2024 13:03 |
|
PerrineClostermann posted:So what's up with these new Bay Trail Atoms? Why are they so damned good? The architecture of the old atom was really antiquated, which is why its performance was never anything to write home about. The bay trail has a completely revamped, more modern architecture which results in a huge performance boost.
|
# ¿ Jan 6, 2014 00:08 |
|
PerrineClostermann posted:What architecture was the old Atom based on? And the new one? The old architecture was Bonnell and the new one is Silvermont. The biggest factor in the performance difference is that the Bonnell had in-order execution and the Silvermont has fully out-of-order execution and improved branch prediction, which results in like a 50% or more increase in the amount of instructions it can execute in a single clock cycle.
|
# ¿ Jan 6, 2014 00:33 |
|
Yeah it sucks to know that even though there's tons of desktop enthusiasts pining to upgrade their sandy bridge, we're just too small and too niche of a market for anyone to give a gently caress anymore. Haswell-E will provide what you want (more cores for under 4 figures) but I'm assuming it will basically just be a 6-core version of the 4770k for $600 and a 8-core version for $1000.
|
# ¿ Feb 22, 2014 23:08 |
|
wheez the roux posted:What kind of mental gymnastics does it take to consider this anything but a positive? Have you ever, even once in the past decade, been limited in any capacity whatsoever by CPU speed? Well if you were the sort of user that didn't need big CPU processing power nothing was forcing you to upgrade, it's just that now the people who do need that power don't have any upgrade options.
|
# ¿ May 27, 2014 15:49 |
|
staticman posted:DC is giving me a real bad PC building itch. I'm on an i7 960, base clock, and was originally planning to wait until Broadwell as well. Since my CPU is a quite a bit older than your Ivy Bridge, I'm curious if I'm a complete dumbass for sticking with it. That's still not that terrible of a CPU despite its age, Devil's Canyon might actually be better than Broadwell though or at least no worse so I think it would be a prime time to upgrade.
|
# ¿ Jun 5, 2014 01:42 |
|
This is the first full review of the i7-4790k I have found, lets hope that this guy was just unlucky and that this isn't typical. While it ran cooler than an i7-4770k it had very little additional overclocking headroom. http://www.pcper.com/reviews/Processors/Intel-Core-i7-4790K-Devils-Canyon-Review-and-Overclocking
|
# ¿ Jun 7, 2014 19:38 |
|
Welmu posted:I've a few questions: Yeah, if they were able to manufacture every chip as a mirror copy of that chip able to hit 5.5GHz then they would all have that capability. quote:- Thanks to the rapid cadence of Intel's tick-tock, does electromigration really matter all that much? If I'm going to buy a new processor anyway in 2-4 years why wouldn't I run it into the ground during that time? People who pimp their cars with nitrous injection and whatnot probably care more about performance than resale value They're still doing the tick-tock but it's not resulting in the same sort of gains that it used to. Sandy Bridge is well over 3 years old now yet it's still basically neck and neck with Haswell when overclocked because it has more overclocking headroom. That said, it's not likely that you're going to burn out your processor with a modest overclock even if you do keep it for a couple years longer due to the slowing pace of performance increases. quote:- How much of overclocking is actually beneficial and how much is ePeen? Does raising CPU clocks by ~one gigahertz really make a system that much faster if the rest of the components are sound (i.e. 1866 dual-channel RAM, SSD, discrete GPU to help with video encoding)? If the task is CPU intensive overclocking will definitely help, if you can get 1GHz of real overclock that would certainly make a difference. The point isn't that "overclocking is epeen" it's that overclocking is slowly being phased out essentially. Turbo boost narrows the gap between stock clocks and what you'd get while overclocked. Newer chips are being sold with less headroom than they used to have. Smaller process nodes have voltage/power consumption characteristics that are not amenable to overclocking. When you feed these chips more voltage their power usage and heat ramps up faster than they used to on larger nodes. All of these factors contribute to make overclocking less appealing from a price/performance standpoint. When you could buy a $30 cooler and overclock your Sandy Bridge from 3.4 to 5GHz that was definitely not epeen and actually added real value to the system, but those days are probably permanently over. MaxxBot fucked around with this message at 17:45 on Jun 8, 2014 |
# ¿ Jun 8, 2014 17:43 |
|
Shaocaholica posted:I don't even remember how to program FPGAs from college but that was soooo long ago and its probably not even close to this thing. Is there a compiler or language? Seems like there should be some high(er) level programming for something this powerful. The two main languages are VHDL and Verilog, with the latter being more popular I believe. I've done a fair amount of Verilog and it's not too bad, the syntax is sort of similar to C.
|
# ¿ Jun 20, 2014 01:01 |
|
Krailor posted:The only people that the Pentium is really good for are serious OC enthusiasts who are planning on doing suicide runs for e-peen. That way they only destroy a $75 processor vs. $200-$300 for an i5k/i7k. It has a 50W TDP, I don't think you'll need to spend that much on a cooler to get a decent overclock.
|
# ¿ Jun 27, 2014 20:40 |
|
Swartz posted:I have an i5 2500k @ 4.2ghz. Would you say it's worth it to upgrade to Devil's Canyon or wait to upgrade to Broadwell or Skylake? I use it primarily for gaming but also 3d modelling. Even as someone obsessed with having the fastest hardware out there I decided against upgrading from my 2600k to Devil's Canyon, it's just not a big enough performance increase. You might be interested in Haswell-E if improving 3d modeling performance is a high priority otherwise I'd just wait.
|
# ¿ Jul 24, 2014 20:50 |
|
Agreed posted:I just wish I knew how much of that was a lack of market pressure, and how much of it is "no seriously processors are loving HARD TO MAKE." I think it's a degree of both, and the loving hard to make part applies not just to the actual fabrication but also to the microarchitecture. By that I mean that while Intel is still able to keep cramming more transistors on the chip when they want to, there's also the challenge of making those extra transistors translate into meaningful performance gains. With GPUs and server CPUs things are a bit easier because you can just keep using the MOAR CORES approach at least at the current time, while with desktop CPUs most software doesn't make that approach especially useful. There's also the fact that Amdahl's law puts theoretical limits on that approach in many situations even when the software does catch up to increasing core counts. When more cores doesn't help the only other option is to increase single threaded performance, either by increasing the clock frequency or increasing IPC. The clock frequency gains slowed to a crawl a decade ago so really the only approach that leaves is the IPC approach. To increase IPC you add more cache, more associativity, more advanced branch prediction, more registers, more execution ports, etc but that stuff is all starting to reach diminishing returns AFAIK, meaning that you don't get the same performance increases per amount of added transistors that you used to be able to. There are limits on exploiting instruction level parallelism meaning that just adding more parallel units to the pipeline isn't particularly useful after a while because you'll never keep them all fed. That said, if Intel has actual competition in the desktop CPU market and there was more demand I'm sure they'd find a way to provide more than the piddly performance increases they now offer. They just wouldn't be able to offer large performance jumps every generation such as with GPUs for example because Agreed posted:processors are loving HARD TO MAKE
|
# ¿ Aug 14, 2014 19:10 |
|
Sidesaddle Cavalry posted:Yeah, there are some good ideas here for how to get new tech onto a smaller form factor, but the practicality of working with a 140W processor in an enclosed ITX case is going to be a sticking point. What specific applications for this platform will need to be that small? I don't need X99 but if I did I can see the appeal of wanting to use an mITX case. I'm just glad they released this product because I had seen a lot of people bring up the concept of mITX X99 before and it was always shot down as physically impossible, which it clearly is not.
|
# ¿ Mar 15, 2015 18:46 |
|
Tab8715 posted:What's Zen? Unless that 2500k is sitting at 5GHz or something you would get more than 10%, the benchmarks I saw were showing around a 8-10% increase from Haswell. MaxxBot fucked around with this message at 00:57 on May 30, 2015 |
# ¿ May 30, 2015 00:55 |
|
HalloKitty posted:In games, at least, the difference isn't impressive. Yeah, I should have specified that I was referring more to performance in multi-threaded CPU intensive tasks rather than games. I don't think we'll ever see a CPU give a big boost in gaming again for various different reasons. Col.Kiwi posted:That person has a seriously above-average chip. Yours I think is more average. I "sadly" have a below average 2500k in that it will NOT go beyond 4.2 and be stable, no matter what. But a 4.2GHz 2500k is still pretty goddamn good to be quite honest. It's definitely above average but quite a few people were hitting those numbers, I had my 2600k at 5GHz on air but for some reason it runs hotter now than it used to so I can't crank the voltage up that high anymore. Still not bad having it at 4.5GHz on a $20 air cooler though. MaxxBot fucked around with this message at 19:12 on May 30, 2015 |
# ¿ May 30, 2015 19:10 |
|
If desktop users were a big enough market they would probably focus more on products for us but it makes more sense economically for them to focus on servers and mobile. Extracting more single-threaded performance at this point is difficult and expensive, so combined with the shrinking desktop market they don't have the financial incentive to really put a lot of effort into it.
|
# ¿ Jun 2, 2015 17:23 |
|
Lord Windy posted:Does single threaded performance need to get much better than the 5~10% we are getting? I figured it was us that needed to get better with utilising cores. It's not that we don't need more it's that we're not likely to get more than that at this point, clock speeds can't be pushed much higher and the various methods we have used to exploit more IPC over the years are starting to hit diminishing returns. It's much easier to just increase core counts rather than try to further increase single-threaded performance after a certain point, and as people pointed out before most server or HPC applications can take advantage of the more cores approach.
|
# ¿ Jun 3, 2015 01:55 |
|
Theoretically ARM has the advantage of the decoder logic taking up a smaller portion of the CPU die but there's also been a trend over time of the decoder logic taking up a smaller portion of the CPU die across all microarchitectures so this might not end up mattering that much. From what I have seen so far Intel still has a big advantage in performance/watt due to their superior fab when compared to a theoretical ARM competitor, especially with their low power Xeons. It would be interesting to see how the performance/watt would compare if Intel made some ARM cores, which it looks like might happen before too long. EDIT: But fundamentally saying "Intel's going to be eaten alive by ARM" doesn't really make any sense. ARM is an ISA and Intel is a chip manufacturer, if Intel eventually sees an advantage in dumping x86 for ARM they will do so and go along designing and manufacturing chips just as they did before. http://seekingalpha.com/article/3229806-intel-becomes-an-arm-chip-maker MaxxBot fucked around with this message at 04:39 on Jun 3, 2015 |
# ¿ Jun 3, 2015 04:22 |
|
frunksock posted:I'm disappointed. Not only by Skylake, but also the quality of that Anandtech review. What's wrong with it? Just curious, they seem to still be significantly above average as far as PC hardware review sites go.
|
# ¿ Aug 9, 2015 18:59 |
|
WhyteRyce posted:Intel trying to cram Skylake into a whale phone That actually would but only if it runs a full desktop OS and I have doubts that will be the case.
|
# ¿ Sep 2, 2015 18:43 |
|
Yudo posted:Except most desktop chips are loving terrible, leaky rejects as it is and have features fused off, not binned. The extra SMT flag and state registers on an i5 work just fine, but they have been disabled, for example. PC enthusiasts are a tiny proportion of that $27 billion though, most of it is corporate and people who will just buy the new Intel CPU because it's the new Intel CPU rather than because it's X% faster in benchmarks. People who actually need compute performance for professional work buy Xeons which are still seeing sizable performance increases every generation. PC enthusiasts are a niche market and Intel, seeing that they have no competition in the area, made the financial decision to not really bother trying to cater to them anymore.
|
# ¿ Sep 3, 2015 18:29 |
|
Are there any good reviews out of the new mobile Skylake chips with benchmarks comparing them to Broadwell and Haswell? I've found some reviews of laptops with Skylake chips but none with good benchmarks or comparisons.
|
# ¿ Nov 23, 2015 05:49 |
|
canyoneer posted:I thought that more cache = bigger die = way more expensive? Yeah, cache takes up most of the die area on modern desktop CPU cores and I'd imagine a good chunk of the power as well. I don't think they were saying that a gig of cache is technically feasible just that it would be a straightforward (in theory) way to increase single-threaded performance.
|
# ¿ Nov 23, 2015 23:57 |
|
EdEddnEddy posted:I've had mine for 4 I've had my 2600k heavily overclocked and overvolted and often running at 100℅ load for 4 1/2 years now. Originally I was able to run at 4.9GHz but I've backed it down to 4.5GHz, mostly because my cooler doesn't seem to work as well as it used to. Normally I'd upgrade sooner but I want something that's actually faster, I'll probably go with Broadwell-E. MaxxBot fucked around with this message at 00:09 on Dec 5, 2015 |
# ¿ Dec 5, 2015 00:04 |
|
Boiled Water posted:A while back a generous comp sci goon effortposted about Xeon Phi and how it was hot garbage. Does said goon have a comment? I don't know what their complaint was but the old Knight's Corner Phi and the new Knight's Landing Phi are a lot different. The old Phi's cores didn't have the greatest performance, were not x86, and as a result the Phi was only available as an AIC and could not be used as a main CPU. The new Phi has much faster cores that are x86 and is available as a main CPU, which all together should make it a lot faster and easier to program than the old Phi was.
|
# ¿ Mar 29, 2016 22:59 |
|
No Gravitas posted:Given the cash, just buy a few of the natex deal I linked above, it will at least run everything you throw at it. I am kind of tempted, how is the motherboard availability for those?
|
# ¿ Mar 30, 2016 01:04 |
|
BIG HEADLINE posted:As I recall, the rumor about that was that it was a ~speshul~ chip just for the NSA. But I think the original rumor was outed by WCCFTech, which should tell you all you need to know. Broadwell-E is supposed to have a 10-core but it will probably be $999. I'm hoping for a ~$500 8-core for my next system.
|
# ¿ Apr 1, 2016 04:22 |
|
Prescription Combs posted:That new NUC is sweeeet. This is major overkill for what I'm going to be using it for (basically simple HTPC stuff) but I couldn't resist, can't wait to get it. Something about having such a powerful machine in a tiny little box really makes it appealing to me. MaxxBot fucked around with this message at 22:07 on Apr 10, 2016 |
# ¿ Apr 10, 2016 22:00 |
|
Atomizer posted:Oh I know about the SSDs in general, I'm just asking about those parts in particular as a fantasy purchase; I didn't get the NUC preorder in time so I'm not buying any components yet. That Intel 540s is a new, low-end TLC line; Intel in name only. That's why I said it's probably behind the SanDisk in the 1 TB m.2 category. There's supposed to be a 1TB version of the 950 Pro coming out before too long, I'm just gonna get the 512GB model and leave the other M.2 slot free until I run out of space and there are new drives on the market.
|
# ¿ Apr 11, 2016 22:21 |
|
VulgarandStupid posted:I just think it's hard to justify a usage case for it. It has low-middling gpu power, but relatively high processing power. I think most people would be better off with a laptop + Razer Core than the NUC, you get more utility out of that. But realistically I think most people would be better off with a cheaper laptop plus a gaming PC, considering what a Razer core costs. Well as you said given that it has more CPU power than GPU power it makes more sense as a tiny but powerful workstation than as a tiny gaming box. That's definitely a niche item but a NUC that costs at least $900 for a full build is always going to be a niche item. I got it because I wanted a full power PC that could sit unobtrusively under my TV, it's less than 1/10 the volume of a small mini ITX computer. The place I ordered mine from doesn't have tracking so I have no idea if mine shipped yet or not . MaxxBot fucked around with this message at 23:43 on May 13, 2016 |
# ¿ May 13, 2016 23:35 |
|
blowfish posted:90% of laptops are glorified facebook machines that occasionally have to run office. That's what all manufacturers target. blowfish posted:The typical consumer has like 5 tabs open at most. I don't know what proportion of laptops are used as corporate machines but I see a ton of people around the office with hilarious amounts of stuff open at once, including non tech-savvy people.
|
# ¿ May 20, 2016 21:30 |
|
BIG HEADLINE posted:If the *entry-level* processor for that socket is less than $600 at launch I'll be shocked. *starts saving up* At this point if they did release something like that I'd expect them to charge like $5000.
|
# ¿ Jun 1, 2016 15:49 |
|
Potato Salad posted:Finding prime numbers of 2n-1 form? Prime95 uses FFTs to implement large integer multiplication and FFTs are used in a wide variety of scientific and engineering applications.
|
# ¿ Jun 9, 2016 19:05 |
|
For some reason corporate PC use never comes up in these types of conversations, I see people in the office with mountains of programs open at once all the time, even people who aren't doing anything technical. Everyone in the office has dual monitors now too which further encourages having more stuff open.
|
# ¿ Aug 30, 2016 20:49 |
|
karoshi posted:Like, f.ex., in-order GPUs running 4000 threads on a single die. There was a marketing presentation by NV* comparing the FLOPs/rest-of-it power ratio for a GPU and a OOOE x86 core, IIRC it was around 2 orders of magnitude. The OOOE machinery is complicated, remove it and you go from 2 IPC to 10 CPI, add 20 threads and you're back at 2 IPC with less power overall (numbers pulled out of my rear end). Also important in making that high FLOPS/W figure possible is that GPUs and other throughput optimized processors have a lot less cache per core than traditional CPUs which makes a big difference because big caches use a ton of power and area. Big caches are needed to run one thread really quickly but if that's not important you can make a lot more efficient use of the available die and power budget. MaxxBot fucked around with this message at 17:43 on Dec 20, 2016 |
# ¿ Dec 20, 2016 17:38 |
|
PerrineClostermann posted:And seems to become more like SSDs/Standard NAND than DRAM by the day. Rastor posted:All our memristor dreams, like dust in the wind. It seems waaaaaay to early to make these sorts of judgments, at least wait until there are some products out.
|
# ¿ Jan 3, 2017 22:01 |
|
Seems like ARM server could be really badass if someone put down the $$$ to make a chip on a newer process. It seems like the current ARM server stuff is all on older fab tech probably for cost reasons and therefor is going to have a hell of a time trying to compete directly with Intel's latest on performance/watt and such.
|
# ¿ Jan 10, 2017 20:35 |
|
PerrineClostermann posted:The marketing numbers for performance are now under a magnitude of difference between it and traditional SSDs, right? Optane is looking less like a miracle all the time. I think the really big marketing numbers like the "1000x" stuff was based on the speed of the Optane itself vs the speed of NAND. The smaller speedup numbers they've shown off more recently is the speed of an Optane SSD vs an NVMe NAND SSD.
|
# ¿ Jan 18, 2017 19:52 |
|
What kind of marketing voodoo is this chart based on? 6th and 7th generation are pretty much the exact same chip with identical IPC, are they assuming 15% higher clocks?
|
# ¿ Feb 10, 2017 01:02 |
|
|
# ¿ Apr 23, 2024 13:03 |
|
Combat Pretzel posted:If this were true, this'd go over so goddamn well (like not). While most people just browse the drat web, a fair share of them have very occasional bouts of power usage, and this is where ARM will gently caress them over and upset them. Both in performance and app availability. Actually Apple's custom ARM designs are really good and I have little doubt that they could make good performing laptop or desktop ARM chips. It's probably mainly the software reasons that have prevented them from doing this.
|
# ¿ Feb 13, 2017 08:23 |