|
Hey BobHoward - apparently I may have bought a M0 stepping with that E5-2650v3 ES. My CPUID string is showing as "0x000306F2" which shows as the M1 stepping in the specs I can pull... but CPU-Z shows it as a "R2" stepping but with an "ES" in the name. Maybe a QS? Any idea what I've got here? I'm too lazy to pull the cooler back off and try and read the chip. If it's M0, any chance you have errata for it? edit: I'm wondering if the fact that I booted Ubuntu and installed the intel-microcode package might have changed up what CPU-Z is reading the stepping as. That might possibly be bad (improper microcode?) given that it apparently read as the M0 stepping before - but at least it's still reading as "stepping 2" even though the release ID/CPUID value has changed. Is this like "if it boots don't worry about it" territory or are there subtler errors I should worry about here? Apparently my mobo does support the M0 stepping so I should be OK, right? Paul MaudDib fucked around with this message at 08:04 on May 3, 2017 |
#
?
May 3, 2017 04:42
|
|
|
|
| # ? Apr 19, 2024 02:48 |
|
|
I don't know much about that offhand, TBH. If it is an ES it may appear to boot fine yet still have unpleasant bugs that you might not want to live with. Kinda doubt Intel supports ES silicon after general release, so any errata which were fixed in production steppings aren't likely to get any workarounds in microcode updates, operating systems, or whatever.
|
|
#
?
May 3, 2017 09:22
|
|
|
More details about the Xeon Gold / Xeon Platinum stuff: http://www.anandtech.com/show/11332/intel-unveils-the-xeon-scalable-processor-family-skylakesp-in-bronze-silver-gold-and-platinum 
|
|
#
?
May 4, 2017 15:50
|
|
|
I guess it's a good marketing move when you need to convince your manager not to buy pleb tier chips.
|
|
#
?
May 4, 2017 17:16
|
|
|
Boiled Water posted:This is correct. Turns out it's only Q-series (?) boards are affected, and those live in servers. It's still an issue, but slightly smaller than semi accurate wants to make out. q series are generally not server but boring business workstations, at my work we have a bunch of Dell optiplex 7020 and hp elitedesk 800 's that are q chipsets
|
|
#
?
May 5, 2017 03:05
|
|
|
HP 8000 and 8200s have Q chipsets.. is there a way to fix this remotely? There is absolutely no way I can run bios updates on hundreds of desktops across 40 offices
|
|
#
?
May 5, 2017 03:12
|
|
|
mewse posted:HP 8000 and 8200s have Q chipsets.. is there a way to fix this remotely? There is absolutely no way I can run bios updates on hundreds of desktops across 40 offices https://downloadcenter.intel.com/download/26754 You will need remote Windows command-line access or Remote Desktop.
|
|
#
?
May 5, 2017 03:59
|
|
|
Double Punctuation posted:https://downloadcenter.intel.com/download/26754 Thanks
|
|
#
?
May 5, 2017 04:01
|
|
|
Boiled Water posted:I guess it's a good marketing move when you need to convince your manager not to buy pleb tier chips. It seems to have been originally conceived to better mark out chips based on features, not core count. So, a company that wanted the best AES performance in a platform accelerator, for instance, would previously have bought e7 class chips, even though an e3 with the same AES engine enabled would be better because it had a faster base and turbo clock, as the platform isn't dual socket or high memory. It's currently a bit jacked up, because it was pasted over the existing product stack, and marketing got their hands on it. I wonder if intel can make it stick or not. Their target market is really OEM's, not customers, and OEM's have resisted such changes in the past.
|
|
#
?
May 5, 2017 04:16
|
|
|
Uh. http://www.theregister.co.uk/2017/05/04/intel_i77700_heat_spike_problems/
|
|
#
?
May 6, 2017 06:32
|
|
|
SwissArmyDruid posted:Uh. Pretty obviously a TIM problem. There's been a background murmur about this since launch but it's blown up today for some reason. In most cases it gets worse over time, presumably as the TIM migrates out of contact with the surfaces or whatever. Heat often makes various things less viscous so the heavy overclocking that many 7700Ks undergo might be a factor as well. I don't get why Intel keeps loving this up. First Ivy Bridge, then Haswell was a bit better, finally they licked the problem with Devil's Canyon. I guess I don't know about Broadwell (literally nobody owns it* and there's a tiny sample size), but Skylake's TIM was solid but not amazing and then Kabylake is all over the place again. What exactly was wrong with the Devil's Canyon TIM that Intel felt compelled to replace it? * which is sad because the eDRAM really helped minimum frametimes - in many games the 5775C's frametimes are faster than a 6700K. I very much wonder if that's not going to be Intel's short-term reaction to Ryzen - their average frametimes are already great and adding eDRAM again would fix the minimum-framerate problems they have vs Ryzen. Paul MaudDib fucked around with this message at 07:11 on May 6, 2017 |
|
#
?
May 6, 2017 07:04
|
|
|
more likely they put in miniature-navi into an intel chip with edram as their new hot laptop cpu line imo
|
|
#
?
May 6, 2017 09:32
|
|
|
It definitely seems like something is off with the 7700k. I just got one a week or so ago (first time I've ever just bought the high-end chip) and holy gently caress does it spike all the time. I actually even broke down and got an AIO watercooler for it and things seem much better though it'll still spike to 90 occasionally.
|
|
#
?
May 6, 2017 12:49
|
|
|
Paul MaudDib posted:Pretty obviously a TIM problem. There's been a background murmur about this since launch but it's blown up today for some reason. In most cases it gets worse over time, presumably as the TIM migrates out of contact with the surfaces or whatever. Heat often makes various things less viscous so the heavy overclocking that many 7700Ks undergo might be a factor as well. First time I've heard of this problem but it should be easy to check by observing the behavior, delidding, applying new TIM and checking if it stops. My 2013 Crystalwell rMBP had significant thermal issues with massive throttling (50% of nominal clockspeed on all four cores under load). I replaced the TIM which was a was a dry and crumbly mess and clock frequencies went back to 130% of nominal clockspeed at full load on all four cores.  As you mentioned the constant heat does deteriorate thermal interface materials over time and OEM pastes are often much worse than what you can get for aftermarket OC purposes. The case of my rMBP has a whiff off planned obsolescence about it but that's  -material. Did I mention that the newest MBPs have their chips and heatsinks attached to the rear of the PCB and repasting requires the whole logic board to be removed?
|
|
#
?
May 6, 2017 13:11
|
|
|
I thought the TIM wasn't a factor, since people who delidded still had the temp-spike problem, according to the article?
|
|
#
?
May 6, 2017 13:16
|
|
|
JnnyThndrs posted:I thought the TIM wasn't a factor, since people who delidded still had the temp-spike problem, according to the article? Huh yeah, I missed that part. Based on some user reports it spikes much faster than it should be possible given the specific heat capacity of a CPU cooler. If it isn't a TIM issue then perhaps a thermal sensor issue.   (this is with watercooling. source) eames fucked around with this message at 14:03 on May 6, 2017 |
|
#
?
May 6, 2017 14:00
|
|
|
I am not very happy with my new 7700k build either. I bought a H100i v2 and the temps were too high. So I bought a delidding kit and put the Coollaboratory Liquid Ultra on the die, and relidded it. That brought the temps down some, but the last key to the puzzle was that Asus AI3 tool was stepping on my CPU fan voltage after doing all of the testing it does. So I had to create custom fan profiles to change the CPU fan to 100% all the time. Also, the fans that come with the H100i v2 are really loving load, and the load spikes they are talking about made them super annoying. I replaced them with Nocuta NF-F12 and I am much happier, but those fans cost more than half of the AIO to begin with. I definitely see the temp spikes, although they were somewhat mitigated by replacing the TIM and sorting out the cooling of my board.
|
|
#
?
May 6, 2017 15:05
|
|
|
I too have a 7700K. IDLE at 4.4ghz (performance mode) is around 35-40C. I don't run Prime 95 because that poo poo isn't useful and these days you can gently caress your poo poo up overclocked. I did notice that on Balanced mode, just sitting at the desktop, the cores jump from like 800mhz to 4400 instantly and all the time.. when that happens temps shoot up a bit. Oddly it seems the temps do that less when you are stuck at max (4400Mhz). Not watercooled. Just a 30 bux Coolermaster EVO. Totally silent. *shrug*  redeyes fucked around with this message at 16:02 on May 6, 2017 |
|
#
?
May 6, 2017 15:21
|
|
|
I am running at 4.8GHz with an offset of -2 for an effective 4.6GHz under heavy AVX usage, which was also a major step to reducing my thermals. Before delidding, I'd hit 80-85 easily, from an ambient temperature of about 24 degrees. After delidding, fixing fan settings, and running the case fans a little faster to make sure I have good intakes, gaming has me around 65-70 and full Handbrake load tops out around 75 degrees, but by no means is it quiet. At idle, basic browsing ,and sysadmin work, I'm in the 30's. A system load of 10-15% will make my fans spin up, which is annoying. This mostly comes as I do video conferencing / webex for work, where web video shouldn't be stressing a high end CPU at all. The cores spike which triggers the fans, which really don't need to run for such a brief spike.
|
|
#
?
May 6, 2017 15:50
|
|
|
Dell, Lenovo, and HP have advised they'll be releasing updated AMT firmware for CVE-2017-5689 in the next couple of weeks. Some of the machines receiving updates are more than 6 years old.
Mr Chips fucked around with this message at 02:12 on May 7, 2017 |
|
#
?
May 7, 2017 02:06
|
|
|
I have one Intel board that is affected but its end-of-life, not sure if Intel will ever give me an update.
|
|
#
?
May 7, 2017 02:27
|
|
|
God why is there still no leaks about x299 Ive got a burning urge to do a new full retard pc build
|
|
#
?
May 7, 2017 04:27
|
|
|
I thought everything relevant had been leaked about that chipset though? Honestly seems a bit blah to me and those 4 core chips that are being released for it are flat out dumb. Purley looks a hell of a lot more interesting feature-wise but probably going to be too expensive to be practical and doesn't seem to really be targeted for any HEDT stuff. Chipset integrated 10Gbe would be cool to have though.
|
|
#
?
May 7, 2017 04:38
|
|
|
It seems pretty clear the 4-core chips are there so high-end users (and wannabes) can buy a 4C8T and be on the bottom end of the HEDT platform instead of the very tippy-top of the consumer platform. Then you can tell yourself that when you hit the lottery you'll definitely buy a 12-core CPU. Unless the pricing structure changes I think the 6-core chips are still the obvious value leader though, and if Intel actually puts them on the consumer platform (I'm not sure that's possible but OK) then it'll be a pretty solid counter to Ryzen given Intel's continued dominance in single-thread performance. Given the choice I'd rather buy the 6-core on the HEDT platform and get the quad-channel memory and extra PCIe lanes and all that other poo poo though. The pricing's really not that different on the motherboards. Some price cuts on the 8C would be real nice though, maybe like $600. And get rid of the 6850K equivalent and just push 40-lanes down to the 6C SKU, while I'm dreaming. Paul MaudDib fucked around with this message at 04:52 on May 7, 2017 |
|
#
?
May 7, 2017 04:47
|
|
|
Yeah I think that is the goal with those 4C chips too but I don't really see too many biting unless Intel really does make them actually cheap. Basin Falls platform feature wise looks like a good counter to the 8C16T and 6C12T Ryzen's but I wonder how its going to look vs AMD's X399 platfrom. THAT is what we really have little info on so far and if AMD really does release reasonably priced 10C20T or 16C32T chips at OK to good clocks then I don't see how going for a Basin Falls based system is going to make much sense for HEDT enthusiasts. Yeah the Intel brand still carries plenty cachet there but mo' cores=mo' bettah could easily over ride that level of thinking too. I'm sure people who actually need that many cores would be absolutely thrilled too of course.
|
|
#
?
May 7, 2017 05:00
|
|
|
Is Skylake-X still a single die though? Because to be real honest Ryzen's cross-CCX latency already sucks in a 2-wide configuration and going to a 4-wide configuration ain't gonna make it any better that's for sure. I really think Naples-based gaming builds may turn out a lot worse than people think they will. Games scale well up to at least 6 cores, they scale OK to 8 cores (AMD did a nice job making this argument), and the overkill crowd may want to buy 10 and 12 cores, but they don't scale infinitely. 16 cores isn't going to do poo poo for gaming. Intel is still kicking rear end on a core-for-core basis and given that there's a very practical limit to the number of cores you can use in gaming... Skylake-X is still gonna have some buyers. In fact if you believe AMD that "mo' cores" is the future it's a pretty obvious buy since you get both single-thread performance and cores.
|
|
#
?
May 7, 2017 05:11
|
|
|
Seems Intel's going evil again: https://hardforum.com/threads/gamers-red-hot-with-fury-over-intel-core-i7-7700-temperature-spikes.1931553/page-3#post-1042983798 The explicit mention of de-lidding voiding the warranty is a change - prior to recently it was a "we'd prefer you didn't, but if you do, just don't break the chip in the process and we're cool" policy.
|
|
#
?
May 7, 2017 05:15
|
|
|
phosdex posted:I have one Intel board that is affected but its end-of-life, not sure if Intel will ever give me an update. Probably belongs in the secfuck thread, but anyway: https://www.embedi.com/files/white-papers/Silent-Bob-is-Silent.pdf TLDR seems to be that the default admin account password checking was utterly broken
|
|
#
?
May 7, 2017 05:47
|
|
|
Paul MaudDib posted:Is Skylake-X still a single die though? Paul MaudDib posted:Because to be real honest Ryzen's cross-CCX latency already sucks Paul MaudDib posted:16 cores isn't going to do poo poo for gaming. Paul MaudDib posted:Intel is still kicking rear end on a core-for-core basis
|
|
#
?
May 7, 2017 06:02
|
|
|
PC LOAD LETTER posted:Ehhhh they're doing OK, Sandy Bridge vs BD is kicking rear end in my book but whatever, the value perspective is what really is turning a lot of folks off on them from what I see. Yeah, but I mean, an OC'd 5820K comes pretty close to keeping up with an OC'd Ryzen 8C, and that came out like 3 years ago at roughly the same prices as the cheapest Ryzen 7. No value problem there. The 8 core Haswell-Es and Broadwell-Es are still solidly faster than Ryzen 7, the problem there is (as you note) the pricing... but after Skylake-X launches, the 6-cores should pretty much be keeping up with Ryzen 7 and for a given level of multi-thread performance it's always preferable to have it spread across fewer cores (i.e. higher single-thread performance). That goes double for gaming. All Intel really has to do is keep the 6-core Skylake-X chips at the same price, and their improved performance should pretty much neutralize first-gen Ryzen. Like I said the 8-cores do need a price drop but at $600 they would wipe out Ryzen entirely from the high end. And Intel needs to make room at the top of the stack for the (mainstream) 10C and the (X) 12C SKUs. (and again, having 40 lanes even in the lowest-end Skylake-X SKUs would be another solid argument to jump to the HEDT platform over Ryzen, since ideally you'd really want at least 32 lanes for SLI, and they must be right from the processor rather than through the southbridge.) Putting eDRAM back onto the 4C parts should fix up the minimum framerate problems on the 7600K and 7700K, drop the unlocked i3 SKU to $125, and we're done, Ryzen is toast for another year. Hire me Intel 
Paul MaudDib fucked around with this message at 06:19 on May 7, 2017 |
|
#
?
May 7, 2017 06:16
|
|
|
Paul MaudDib posted:Putting eDRAM back onto the 4C parts should fix up the minimum framerate problems on the 7600K and 7700K, drop the unlocked i3 SKU to $125, and we're done, Ryzen is toast for another year. Hire me Intel
|
|
#
?
May 7, 2017 06:20
|
|
|
Put EDRAM on 6c / non HEDT platform 
|
|
#
?
May 7, 2017 06:21
|
|
|
Paul MaudDib posted:Yeah, but I mean, an OC'd 5820K comes pretty close to keeping up with an OC'd Ryzen 8C, and that came out like 3 years ago at roughly the same prices as the cheapest Ryzen 7. No value problem there. Paul MaudDib posted:the 6-cores should pretty much be keeping up with Ryzen 7 and for a given level of multi-thread performance it's always preferable to have it spread across fewer cores (i.e. higher single-thread performance). That goes double for gaming. Paul MaudDib posted:All Intel really has to do is keep the 6-core Skylake-X chips at the same price, and their improved performance should pretty much neutralize first-gen Ryzen. Paul MaudDib posted:(and again, having 40 lanes even in the lowest-end Skylake-X SKUs would be another solid argument to jump to the HEDT platform over Ryzen Paul MaudDib posted:Putting eDRAM back onto the 4C parts should fix up the minimum framerate problems on the 7600K and 7700K, drop the unlocked i3 SKU to $125, and we're done, Ryzen is toast for another year. Hire me Intel
|
|
#
?
May 7, 2017 06:54
|
|
|
PC LOAD LETTER posted:The problem is that you could say the same about the 5820K vs any newer Intel chip too and I don't think they have a good reply for the 6C12T Ryzens yet on a bang vs buck basis either. Also the 5820K's tend to go for around $400 while the 1700 goes for around $320 and comes with a OK HSF. From a value perspective that is a clear cut win. They're still pretty weak in the i5 segment but eDRAM might help shore that up a little bit. The i7s are doing OK versus the R5s, they clock high enough to not be a total writeoff in productivity tasks and IIRC they spank the R5s in gaming. quote:unless those 6C12T SkylakeX's OC as well as the 4C8T ones (probably not, none of the over 4C chips seem to OC well) The big chips overclock fine, they're typically 100-200 MHz below their smaller counterparts. The problem is that Broadwell doesn't overclock well to begin with. In either its small chip or large chip form. The small chips had the eDRAM, the big chips were a big wet fart for the enthusiast market. Bet the server guys liked the power gains, but enthusiasts overclock it anyway. Haswell overclocked fine, and Haswell-E hits 4.5 GHz reliably if you throw voltage at it. quote:Skylake only has like a ~9% IPC advantage over Zen ... they're not very compelling at all from a bang vs buck stand point. Yeah it'll win all the benches but real world use you won't see much difference between the 2 for most games if they both stay around 4-4.5Ghz. Yeah but Skylake has like a 15% clock-rate advantage over Zen too. The 6700K reliably hits 4.7 GHz (a bit higher than Haswell) and I'm betting that Skylake-X reliably hits 4.6 GHz (a bit higher than Haswell-E). The R7 1700 only goes to 3.9 GHz in the vast majority of samples, 4.0 is getting amazingly lucky, the ones hitting 4.1 GHz are 1800Xs and you still need to be lucky. (scale is wrong but it's 1700, 1700X, and 1800X, and note that you really don't want to go past 1.4v unless you have good cooling, and never more than 1.45v) That gives a 6700K at 4.6 GHz a 17% clock rate advantage. Combining the IPC advantage, the clockrate advantage, and a little extra efficiency from being spread over fewer cores... that puts the 6-cores Intels pretty damned close to the magic 33% they'd need to compete with 8-core Ryzens. quote:That sounds like a way for Intel to go out of business. They have much higher operating costs than AMD due to the fabs. AMD has lots of room to cut prices and still make money I believe if Intel wants to go the price war route. Not raising prices isn't the same thing as cutting prices. They already raised Broadwell-E prices a bunch, that was then, they have competition now. But holding the pricing 6-core SKUs steady should be OK in terms of competitiveness, as long as they bump up the performance significantly. The 8-cores still need to come down. $1000 is just too much for that SKU now that they have competition. quote:Over Ryzen sure that might be a deciding factor for a few but over X399? Is that what they're calling the Ryzen HEDT platform? That's kinda confusing. Yeah, if Ryzen HEDT has more lanes that would be good but Ryzen is already pushing up against Intel-level pricing and the HEDT chips are probably gonna start at like, $1000, for like 10C. Intel's actually kind of a bargain in comparison especially given the single-thread performance advantages  I bet clock rates are also gonna be low for the lower-end SKUs. quote:It'd be interesting to see if AMD went the HBM route on the mid to high priced APU's to counter this. If this isn't BS then maybe HBM has matured enough to make that practical. It seems like that's clearly where AMD is aiming, and those SKUs are going to be interesting for sure. I've commented on that before in the other thread I think, I don't know how reputable those prices are, one way or the other. Overall the die and BoM prices do sound ballpark right but it's not like it's my job to know or anything. The fact that both dies are equally priced is a little suspicious but perhaps the price per wafer went down, and all in all who knows on HBM. Paul MaudDib fucked around with this message at 07:57 on May 7, 2017 |
|
#
?
May 7, 2017 07:47
|
|
|
Anime Schoolgirl posted:unlike nvidia, intel insists on making really stupid decisions which is why you will never get a job there places I also do not work: NVIDIA, AMD (I applied at both, given I did GPGPU work for my thesis but I got turned down) GRINDCORE MEGGIDO posted:Put EDRAM on 6c / non HEDT platform Nah gently caress that, put eDRAM on the HEDT platform. 8-core dedicated/binned gaming SKU, 256 MB of eDRAM. (list price: $1500  )
Paul MaudDib fucked around with this message at 07:58 on May 7, 2017 |
|
#
?
May 7, 2017 07:49
|
|
|
Paul MaudDib posted:They're still pretty weak in the i5 segment but eDRAM might help shore that up a little bit. The i7s are doing OK versus the R5s, they clock high enough to not be a total writeoff in productivity tasks and IIRC they spank the R5s in gaming. I, too, game with SLI 1080tis with a 288hz 720p monitor 
|
|
#
?
May 7, 2017 07:59
|
|
|
i still find it funny people are legitimately upset and disappointed that AMD 'only' managed to get within like 15% of Intel's latest offering of a decade long refinement process (thats in turn based on an already established design skeleton!) after only a few years of alpha and beta development on a complete brand spanking new design (unless zen is a refinement of the old k6s or whatever)
|
|
|
#
?
May 7, 2017 08:41
|
|
|
Paul MaudDib posted:The i7s are doing OK versus the R5s, they clock high enough to not be a total writeoff in productivity tasks and IIRC they spank the R5s in gaming. Paul MaudDib posted:The big chips overclock fine, they're typically 100-200 MHz below their smaller counterparts. Paul MaudDib posted:Yeah but Skylake has like a 15% clock-rate advantage over Zen too. The 6700K reliably hits 4.7 GHz (a bit higher than Haswell) and I'm betting that Skylake-X reliably hits 4.6 GHz (a bit higher than Haswell-E). Paul MaudDib posted:The R7 1700 only goes to 3.9 GHz in the vast majority of samples, 4.0 is getting amazingly lucky, the ones hitting 4.1 GHz are 1800Xs. Paul MaudDib posted:Not raising prices isn't the same thing as cutting prices. Paul MaudDib posted:Is that what they're calling the Ryzen HEDT platform? That's kinda confusing. Paul MaudDib posted:Ryzen is already pushing up against Intel-level pricing and the HEDT chips are probably gonna start at like, $1000, for like 10C.
|
|
#
?
May 7, 2017 08:41
|
|
|
Aren't we expecting the R3's to clean up in the budget gaming section? I mean if R7's and R5's are all hitting 3.9ghz, I'd say we could probably expect the R3's to do the same. They still aren't super competitive in single core performance when compared to a OCed i5 or i7, but should be do fine against Pentiums, i3's and non-overclocked i5's and should be even more competitive once you factor in not having to buy a premium motherboard to OC or have faster RAM.
|
|
#
?
May 7, 2017 08:51
|
|
|
|
| # ? Apr 19, 2024 02:48 |
|
|
True, as long as intel don't dump pentium prices or release another anniversary edition
|
|
#
?
May 7, 2017 09:13
|
|