|
It's just a pointless anecdote, but Intel really is experiencing manufacturing woes: I heard from a guy at work today that my new laptop will take a while to come due to Intel supply issues
|
#
?
Nov 27, 2019 16:27
|
|
|
|
| # ? May 20, 2024 12:06 |
|
|
HalloKitty posted:It's just a pointless anecdote, but Intel really is experiencing manufacturing woes: I heard from a guy at work today that my new laptop will take a while to come due to Intel supply issues I ordered a laptop for a new hire end of july who was starting in september. He got it last week. I didn’t really believe the supply woes excuse I thought it was just IT being useless but perhaps I was wrong..
|
|
#
?
Nov 27, 2019 16:31
|
|
|
BangersInMyKnickers posted:In the consumer space, yes the gap isn't that great. In the server market, Intel is only staying afloat with institutional momentum. Epyc chips beat them out on performance and cost pretty much across the board.
|
|
#
?
Nov 27, 2019 16:53
|
|
|
Talking about competitive balance long term, if Intels 7nm actually shows up on schedule (lol) in 2021 or 2022, it is very likely to wreck everyones poo poo. The density numbers on it are ridiculous. More than double the density of Intels 10nm, TSMCs 7nm+, or Samsungs "7nm" EUV (which are roughly equal), and 50% more dense than TSMC "5nm". But TSMCs "5nm" is supposed to be already a year old by then, so AMD really does need to keep their foot on the gas while they can. They do have the problem of being a relatively low profit part on TSMCs nodes though, which means they normally don't get a crack until after Apple. Ditching Glofo is a double edge sword I guess. They wouldn't be here without TSMCs 7nm, but now they have to rely on someone else to supply them... someone who does business with bigger fish than AMD.
|
|
#
?
Nov 27, 2019 18:17
|
|
|
Cygni posted:Talking about competitive balance long term, if Intels 7nm actually shows up on schedule (lol) in 2021 or 2022, it is very likely to wreck everyones poo poo. The density numbers on it are ridiculous. More than double the density of Intels 10nm, TSMCs 7nm+, or Samsungs "7nm" EUV (which are roughly equal), and 50% more dense than TSMC "5nm". Every new process node is the best poo poo ever until it has to go into production 
|
|
#
?
Nov 27, 2019 22:50
|
|
|
OTOH, Intel has a long history of absolute dominance in terms of high performance process tech. Yes, 10nm has been a complete disaster, but I don't think that's reason to assume they've completely lost their edge.
|
|
#
?
Nov 27, 2019 23:07
|
|
|
HalloKitty posted:It's just a pointless anecdote, but Intel really is experiencing manufacturing woes: I heard from a guy at work today that my new laptop will take a while to come due to Intel supply issues Anandtech just wrote about this. https://www.anandtech.com/show/15162/dell-intel-cpu-shortages-worsened-in-q4-premium-commercial-pcs-impacted article posted:Intel experienced production variability in the fourth quarter
|
|
#
?
Nov 27, 2019 23:40
|
|
|
priznat posted:I ordered a laptop for a new hire end of july who was starting in september. He got it last week. I didn’t really believe the supply woes excuse I thought it was just IT being useless but perhaps I was wrong.. we have to wait 1.5 months before dell and ship a laptop, and no one ever got fired for buying intel.
|
|
#
?
Nov 28, 2019 06:49
|
|
|
wargames posted:we have to wait 1.5 months before dell and ship a laptop, and no one ever got fired for buying intel. Hmm, this was dell too. We used to buy our own laptops for the team from a list of options but after an acquisition we have to purchase through servicenow tickets to IT and it is a black hole of responsiveness.
|
|
#
?
Nov 28, 2019 06:54
|
|
|
Mine's HP, for what it's worth
HalloKitty fucked around with this message at 09:13 on Nov 28, 2019 |
|
#
?
Nov 28, 2019 08:39
|
|
|
We just received couple HPE Proliant servers with Xeon Gold 6146 processors which took almost month and a half to deliver. Initial estimate for the processors was 60 days. I just ordered another server. On the original configuration the CPU had normal delivery time, but the 64GB RAM sticks had 60 days. I don't know what RAM shortage we are suffering now. Fortunately we could swap it to different similar model with usual delivery time.
|
|
#
?
Nov 28, 2019 12:51
|
|
|
Been receiving biweekly delivery pipeline updates since March because Dell’s delays are so bad we have to defer new hire start dates; we don’t have computers to give new employees. It’s real out there, folks.
|
|
#
?
Nov 28, 2019 15:51
|
|
|
Serious question: My shop is all Macs for dev machines and has had zero of these issues, and we're ordering pretty big quantities. We're even ordering them through an obtuse ServiceNow corporate process and they show up in less than a week. How is Apple sidestepping this?
|
|
#
?
Nov 28, 2019 15:57
|
|
|
It has been hell to order singles but ordering 50+ has been business as usual for me the last few years. I may have lucked out on the refresh though.
|
|
#
?
Nov 28, 2019 16:02
|
|
|
Twerk from Home posted:How is Apple sidestepping this? Wouldn't surprise me if Apple is getting priority on shipments thanks to whatever deals Intel offers every time Apple threatens to jump ship to another manufacturer.
|
|
#
?
Nov 28, 2019 16:06
|
|
|
isndl posted:Wouldn't surprise me if Apple is getting priority on shipments thanks to whatever deals Intel offers every time Apple threatens to jump ship to another manufacturer. It's likely this.
|
|
#
?
Nov 28, 2019 16:34
|
|
|
eames posted:Anandtech just wrote about this. https://newsroom.intel.com/news/intel-supply-update/
|
|
#
?
Nov 28, 2019 16:36
|
|
|
priznat posted:Hmm, this was dell too. We used to buy our own laptops for the team from a list of options but after an acquisition we have to purchase through servicenow tickets to IT and it is a black hole of responsiveness. I am on the IT side this is just how long it takes.
|
|
#
?
Nov 28, 2019 16:40
|
|
|
Intels capacity problems are probably all related to 10nm being behind schedule, 10nm has at least double the density of 14nm, which means double the chips per wafer. I'm assuming wafer throughput is relatively constant regardless of process node.
|
|
#
?
Nov 28, 2019 16:48
|
|
|
Most of their CPUs are still on 14 nm. I suspect one reason for the shortages is that, thanks to AMD, the average CPU now comes with a lot more cores than they used to, making them larger, meaning Intel get fewer CPUs per wafer.
|
|
#
?
Nov 28, 2019 17:55
|
|
|
What ever happened to the rumors of Samsung fabbing 14nm CPUs for Intel to address the shortages? Guess that never materialized.
|
|
#
?
Nov 28, 2019 23:39
|
|
|
Perplx posted:Intels capacity problems are probably all related to 10nm being behind schedule, 10nm has at least double the density of 14nm, which means double the chips per wafer. I'm assuming wafer throughput is relatively constant regardless of process node. Yep, fabs are set up for so many wafer starts per month, what you put on those wafers is your own business, but smaller chips = high yields. eames posted:What ever happened to the rumors of Samsung fabbing 14nm CPUs for Intel to address the shortages? Guess that never materialized. Even if you bust your rear end setting the masks up, going from "I want the shiny" to "here is the first batch of 50k shinies" is like a 6-12 months long, depending on process node, number of layers, and a million other factors. Methylethylaldehyde fucked around with this message at 00:04 on Nov 29, 2019 |
|
#
?
Nov 29, 2019 00:01
|
|
|
eames posted:What ever happened to the rumors of Samsung fabbing 14nm CPUs for Intel to address the shortages? Guess that never materialized. Turned out to be chipsets rather than CPUs.
|
|
#
?
Nov 29, 2019 00:40
|
|
|
I got a screaming deal on a 9900KS and a Z390 motherboard so looks like i'm building a new rig. Asking in here rather than the PC building thread because I vaguely remember someone having strong opinions about it (Paul Muad'dib?): what RAM should I get? Currently eyeing G.SKILL Trident Z Neo RGB F4-3600C16D-32GTZNC 32 GB RAM
|
|
#
?
Dec 2, 2019 09:10
|
|
|
Generally a set of Samsung b die RAM will clock the highest and have better timings but it’s all a crap shoot which depends on your ram, motherboard and IMC. Most 3200 CL14 ram is die and a fair amount of 3600 CL16 is as well. You can wring out some decent performance on intel with ram overclocking but not as much as ryzen. I’m running 3200 CL14 b die at 3700 CL15 with tightened timings a because my mobo or cpu IMC doesn’t like higher than that. Roughly every 200 MHz evens out with a single increase in CL. This was a decent list I saw on reddit. https://docs.google.com/document/d/1y29lfTkhA-YHUBp4ZRbKu6i7QslhJ3I2SMYIsGYe4Qs/edit I’ve had terrible luck with patriot memory so I won’t buy them but I have had good luck most other brands. If you don’t care about tinkering a nice 3600 CL16 16GB kit or 3200 CL16 if you’re looking at 32GB will basically get your 95% of the way there. I used this guide a some other resources to overclock. https://github.com/integralfx/MemTestHelper/blob/master/DDR4%20OC%20Guide.md B-Mac fucked around with this message at 14:14 on Dec 2, 2019 |
|
#
?
Dec 2, 2019 14:08
|
|
|
Sphyre posted:I got a screaming deal on a 9900KS and a Z390 motherboard so looks like i'm building a new rig. Asking in here rather than the PC building thread because I vaguely remember someone having strong opinions about it (Paul Muad'dib?): what RAM should I get? Currently eyeing G.SKILL Trident Z Neo RGB F4-3600C16D-32GTZNC 32 GB RAM If you're not the type to spend hours trying to get the last 5% of performance out of your RAM, that G.Skill kit should be perfectly fine.
|
|
#
?
Dec 2, 2019 14:32
|
|
|
Chrome v79 on my Asus Chromebook just disabled hyperthreading due to those Intel vulnerabilities. I was able to 'temporarily' turn HT back on but wow, it was slow as poo poo without it.
|
|
#
?
Dec 2, 2019 21:18
|
|
|
Dog Toothbrush posted:Chrome v79 on my Asus Chromebook just disabled hyperthreading due to those Intel vulnerabilities. I was able to 'temporarily' turn HT back on but wow, it was slow as poo poo without it. Yeah, I mentioned this a few times over in the Mac thread. All the dualcore 1.6 GHz Macbook Airs that are sold as new and up-to-date are in for a rough ride if (when?) Apple has to disable HT by default.
|
|
#
?
Dec 2, 2019 21:24
|
|
|
Does B-die really hold any advantage over that new stuff (Micron E-die I believe)? The new stuff is a lot cheaper and seems to support some fast frequencies.
|
|
#
?
Dec 2, 2019 22:01
|
|
|
eames posted:Yeah, I mentioned this a few times over in the Mac thread. All the dualcore 1.6 GHz Macbook Airs that are sold as new and up-to-date are in for a rough ride if (when?) Apple has to disable HT by default. Won't i5/i7 processors be affected by this too? I'm not a computer genius but I watched some YouTubers do tests on high end gaming rigs and disabling HT seemed to make a difference.
|
|
#
?
Dec 2, 2019 22:29
|
|
|
Paul MaudDib posted:Does B-die really hold any advantage over that new stuff (Micron E-die I believe)? The new stuff is a lot cheaper and seems to support some fast frequencies.
|
|
#
?
Dec 2, 2019 23:04
|
|
|
Dog Toothbrush posted:Won't i5/i7 processors be affected by this too? I'm not a computer genius but I watched some YouTubers do tests on high end gaming rigs and disabling HT seemed to make a difference. A bunch of laptop i5/i7 are dual-cores. Anything 2C/4T will get his particularly hard I'd bet, as 2 threads just isn't great for modern workloads.
|
|
#
?
Dec 2, 2019 23:08
|
|
|
aren't there utilities that let you disable the security mitigations?
|
|
#
?
Dec 2, 2019 23:09
|
|
|
Dog Toothbrush posted:Won't i5/i7 processors be affected by this too? I'm not a computer genius but I watched some YouTubers do tests on high end gaming rigs and disabling HT seemed to make a difference. Going from 4C/8T to 4C/4T would be a barely noticeable annoyance for the average user (!) whereas going from 2C/4T to 2C/2T will potentially make your computer unusable for a few seconds if your start Mail.app, Safari and Slack at the same time while Time Machine is running. Paul MaudDib posted:Does B-die really hold any advantage over that new stuff (Micron E-die I believe)? The new stuff is a lot cheaper and seems to support some fast frequencies. From what I’ve read it depends on the platform and binning. The good, top end B-die bins still run way tighter timings at the same or higher clocks than Micron E-die, particularly primaries as well as secondaries like TRFC. Intel gets the most performance gain out of these because the IMC scales well and the cache is relatively small compared to Ryzen. B-die also scales better with voltage which probably isn’t important for 24/7 use. Once you go down to the lower bins (sub 3600MHz with mediocre timings) the difference gets much smaller. At that point average E-die sticks may perform better than bad B-die. e: I should also mention that interleaving in dual rank mode generally gives higher throughout but diminishes the value/impact of low timings. So for 2x8Gb on an Apex board you really want B-die while it won’t matter as much for 2x16GB or 4x8GB on other boards. eames fucked around with this message at 00:07 on Dec 3, 2019 |
|
#
?
Dec 2, 2019 23:55
|
|
|
don't those hyperthreading attacks require a lot of involved effort to exploit? why is it worth destroying performance on cheap laptops to protect against an attack that no one outside of a server farm is going to see
|
|
#
?
Dec 3, 2019 00:26
|
|
|
I agree and in a consumer-friendly world there would be a little "opt out of speculation execution exploits" button for consumers. Intel's explanation would probably be that their REAL customers are server farms, and that you should upgrade your hardware anyway. In addition, consider the bad press coverage that would occur if a data breach happened that could be blamed on un-patched operating systems and CPU architecture. mix that up into a stew and you get mandatory performance-limiting software patches.
|
|
#
?
Dec 3, 2019 07:00
|
|
|
Shipon posted:don't those hyperthreading attacks require a lot of involved effort to exploit? why is it worth destroying performance on cheap laptops to protect against an attack that no one outside of a server farm is going to see The concern of the independent researcher teams that found these bugs is that there could be (multiple) undiscovered/undisclosed exploits in the wild that require less effort and/or are harder or even impossible to mitigate.
|
|
#
?
Dec 3, 2019 07:44
|
|
|
Asimov posted:I agree and in a consumer-friendly world there would be a little "opt out of speculation execution exploits" button for consumers. Intel's explanation would probably be that their REAL customers are server farms, and that you should upgrade your hardware anyway. In addition, consider the bad press coverage that would occur if a data breach happened that could be blamed on un-patched operating systems and CPU architecture. mix that up into a stew and you get mandatory performance-limiting software patches.
|
|
#
?
Dec 3, 2019 11:06
|
|
|
Shipon posted:don't those hyperthreading attacks require a lot of involved effort to exploit? why is it worth destroying performance on cheap laptops to protect against an attack that no one outside of a server farm is going to see There's probably some CEO out there where the vector would be absolutely worth using on their laptop and Intel would be liable if they didn't apply a blanket policy
|
|
#
?
Dec 3, 2019 11:44
|
|
|
|
| # ? May 20, 2024 12:06 |
|
|
For what it's worth, the biggest issue with the side-channel attacks was that it allowed javascript running in browsers to be used to leak kernel information in memory including passwords stored in password vaults - and that has been mitigated on both Chrom(e|ium) and Firefox which covers like 98% of the browser market, by making the timings in browsers not as tight. So if you're like me and have an old laptop where the performance impact is rather high (even with the big speed increases in Firefox), you might forego them and rely on the browser mitigation instead.
|
|
#
?
Dec 3, 2019 13:55
|
|