|
movax posted:Is it Alder Lake w/ E cores or just P cores? Have schedulers figured out how to deal with heterogenous cores on x86 yet? Alder Lake-N is all gracemont cores. Aka E cores.
|
# ? Jun 23, 2023 16:50 |
|
|
# ? Apr 28, 2024 06:44 |
|
movax posted:Is it Alder Lake w/ E cores or just P cores? Have schedulers figured out how to deal with heterogenous cores on x86 yet? Just E-cores, if I'm not mistaken. E: beaten
|
# ? Jun 23, 2023 16:51 |
|
Twerk from Home posted:Only E cores, no P cores. The i3-N305 is as far as I know the first Intel Atom lineage part to ever be branded Core i. I thought there was something with Tremont cores but the product I’m misremembering is even weirder than I thought: https://www.intel.com/content/www/us/en/products/sku/202777/intel-core-i5l16g7-processor-4m-cache-up-to-3-0ghz/specifications.html https://www.notebookcheck.net/Intel-Core-i5-L16G7-Processor-Benchmarks-and-Specs.466091.0.html quote:Marketing Status Discontinued
|
# ? Jun 23, 2023 16:57 |
|
hobbesmaster posted:I thought there was something with Tremont cores but the product I’m misremembering is even weirder than I thought: These still show up, the Pentium Gold 8500 is 1 P / 4 E: https://ark.intel.com/content/www/us/en/ark/products/226449/intel-pentium-gold-processor-8500-8m-cache-up-to-4-40-ghz.html It's really wild to me that that gets branded "Pentium" but 8 E cores are "Core i3".
|
# ? Jun 23, 2023 17:24 |
|
Twerk from Home posted:These still show up, the Pentium Gold 8500 is 1 P / 4 E: https://ark.intel.com/content/www/us/en/ark/products/226449/intel-pentium-gold-processor-8500-8m-cache-up-to-4-40-ghz.html Yeah, I meant I thought i5 L16G7 was an atom branded as an i5. But really… what even is that part number. It’s even worse than the Atom x5/x7 for weirdness
|
# ? Jun 23, 2023 17:56 |
|
The wildest intel processor is still the first 10nm chip they made just to comply with a deadline, and they shipped it in a single lenovo laptop in china.
|
# ? Jun 23, 2023 18:46 |
|
HalloKitty posted:1U just isn't a particularly good format for power hungry devices, as you end up wasting a lot of power lost to nothing but turning tiny screamer fans, instead of doing something useful. We have ~24 dual node 1U servers we used for testing some ssds (24 per box). Even long time, experienced data center people stop and say "what the hell is going on there" any time they walk by that aisle.
|
# ? Jun 23, 2023 19:16 |
|
Wild EEPROM posted:The wildest intel processor is still the first 10nm chip they made just to comply with a deadline, and they shipped it in a single lenovo laptop in china. also I believe the only product to ever get released on that initial version of Intel 10nm. rip Cannon Lake https://ark.intel.com/content/www/us/en/ark/products/136863/intel-core-i38121u-processor-4m-cache-up-to-3-20-ghz.html
|
# ? Jun 23, 2023 22:20 |
|
Do we know yet if the raptor lake refresh will be compatible with z690? I'm on a 12600k and it would be rather nice a couple of years from now to just drop a 14700k or whatever into this mobo.
|
# ? Jun 26, 2023 01:00 |
|
Shrimp or Shrimps posted:Do we know yet if the raptor lake refresh will be compatible with z690? I'm on a 12600k and it would be rather nice a couple of years from now to just drop a 14700k or whatever into this mobo. Yes, there are already RPL-R BIOS updates for the majority of Z690 boards. It's definitely coming to everything LGA1700.
|
# ? Jun 26, 2023 06:58 |
|
BurritoJustice posted:Yes, there are already RPL-R BIOS updates for the majority of Z690 boards. It's definitely coming to everything LGA1700. Oh sweet, just checked my Gigabyte board and while their latest bios doesn't specify 14th gen, they do say "Supports and powers up Intel next generation processor". If that's the case then my next upgrade is pretty set, wait for meteor lake to depress prices of the raptor lake refresh and then pop in a 14700k and then just use that for another few years.
|
# ? Jun 26, 2023 07:31 |
|
Judging from the leaks and announcements, there might not be a desktop meteor lake. Raptor lake refresh is supposedly filling that gap.
|
# ? Jun 26, 2023 23:17 |
|
hobbesmaster posted:I thought there was something with Tremont cores but the product I’m misremembering is even weirder than I thought: Lakefield was basically a proving ground for the foveros stacking in Meteor Lake and beyond.
|
# ? Jun 27, 2023 06:33 |
|
Tech Yes City released this video, supposedly part one of a two-part series, that talks about Intel CPUs, starting with the 12th gen, having latency issues: https://www.youtube.com/watch?v=6O5XGVaPDZo the TL; DR based on my understanding: - Intel used to have a Northbridge and a Southbridge chipset, such as with the 1st generation Core architecture - by the 2nd generation, Sandy Bridge, the Northbridge was effectively folded-in with the CPU, and that was a big leap in performance - but, starting with Alder Lake, that functionality was moved off the CPU ... - ... and it's that change that makes 12th/13th gen feel "less snappy" compared to, say, 9th and 10th gen Intel CPUs, when doing things like searching for files or scrolling through a video editing timeline - this was something that the creator only realized/discovered of late, based on discussions with manufacturers during this year's Computex - the effect is comparable to people saying that AMD CPUs, around the time of the Ryzen 3000 series (or earlier) felt "less snappy" compared to their Intel contemporaries, because AMD also did all of their I/O on a separate chiplet from the CPU cores, with silicon that was [still] on GloFo 14nm what's unclear to me is where this I/O functionality was moved to, if it's not where it used to be, relatively speaking, compared to older generations it seems like a rather incredible claim to be making, and I'm not sure what to make of it, which is why I'm bringing it up
|
# ? Jun 28, 2023 08:01 |
|
The guy is full of poo poo. Hes finding Windows 11 bugs and blaming Intel. As if a few more ms of latency will cause massive slowdowns in the GUI. lol.
|
# ? Jun 28, 2023 09:31 |
|
gradenko_2000 posted:it seems like a rather incredible claim to be making, and I'm not sure what to make of it, which is why I'm bringing it up It is, in fact, an incredible claim. I feel dumber after watching about half of this video and I'm not watching the rest. He does not understand enough about this technology to say many coherent things, so I stopped when he finally described his theory of what changed, which is that I/O latency got worse because the System Agent moved off-die. Unfortunately for him, he's full of poo poo, it went nowhere. I'm going to try to explain how this idea really is. You know how modern Intel client CPUs have a very high performance on-die ring bus, most notable for connecting CPU last-layer caches to each other and the memory controller? The System Agent is another ring bus drop, a bridge between the ring and the graphics/DMI PCIe root complexes. It would in fact be a big deal if Intel lost their minds and decided to move the System Agent (or any other ring bus agent) off-die. However, as they are at least minimally competent, they did no such thing. A ring bus agent needs more than 1000 non-SERDES wires at 3+ GHz. Running that through a CPU socket connector is no bueno. There's proof that he's wrong in his own video! He shows an Intel slide with a stylized floorplan of an Alder Lake client CPU. On top, it shows the DMI and PCIe IO blocks. These things would not be on-die if the System Agent was off-die. Also, if his theory is true, there should be quantifiable differences to measure. As far as I was willing to watch, there was nothing but handwaving, and absolutely no sign that this guy can do anything but handwave. His best case excuse for this video is that his poor understanding of CPU and IO tech led him to profoundly misunderstand some comment he overheard at the trade show. BobHoward fucked around with this message at 11:12 on Jun 28, 2023 |
# ? Jun 28, 2023 11:01 |
|
Where exactly was this dumbass seeing everyone saying that the 12gen Intel or Ryzen 3000 were "less snappy", because that is not a thing that happened. And his demos of windows being unresponsive when looking at a bunch of files in explorer has nothing to do with anything. The problem is by default explorer reads files for metadata (id3 tags etc) when looking at music / video folders. At 6:27 is a perfect example, he opens a folder, explorer stalls out, then switches to the media view with different column headers and slowly populates the list. It's annoying as poo poo, but has zero to do with hardware. (The fix is to change to list view in a music/video folder and then do "apply this view to all folders of this type". Or customize the column headers to not be metadata and just be the standard date size & type ones.)
|
# ? Jun 28, 2023 15:30 |
|
My amiga500 has a faster CPU than my current 13gen i7 because it boots to desktop faster.
|
# ? Jun 28, 2023 16:19 |
|
Beef posted:My amiga500 has a faster CPU than my current 13gen i7 because it boots to desktop faster. No lies detected!
|
# ? Jun 28, 2023 16:32 |
|
BobHoward posted:I'm going to try to explain how this idea really is.
|
# ? Jun 28, 2023 19:40 |
|
Klyith posted:Where exactly was this dumbass seeing everyone saying that the 12gen Intel or Ryzen 3000 were "less snappy", because that is not a thing that happened. Hey speaking of this, the lag sucks. I cant seem to make the fix work for me. Dunno why. Is there a reg flag or something to just stop explorer from parsing all this extra crap?
|
# ? Jun 29, 2023 00:04 |
|
redeyes posted:Hey speaking of this, the lag sucks. I cant seem to make the fix work for me. Dunno why. Is there a reg flag or something to just stop explorer from parsing all this extra crap? By default it saves the view settings for every folder you've previously looked at. To clear that, regedit to: HKEY_CURRENT_USER\Software\Classes\Local Settings\Software\Microsoft\Windows\Shell and delete the Bags and BagsMRU sub-folders. Log off/on or reboot. At that point all the media folders will pick up the standard view for their type, which you've set to not read metadata. All the other view customization you've intentionally done are also gone, which sucks, but at least explorer won't take 30 seconds to display a file list. The stupidest thing about this is how they were able to make explorer not poo poo itself over thumbnails. It may take a minute to display them, but you get the file list instantly. But they somehow didn't think to use the same method for metadata.
|
# ? Jun 29, 2023 03:18 |
|
Tech Yes City is a total joke, and this is far from the first time he's tried to stir poo poo about a subject he barely understands.
|
# ? Jun 29, 2023 06:09 |
|
I remember that guy cleaning motherboards with WD40.
|
# ? Jun 29, 2023 12:04 |
|
Dr. Video Games 0031 posted:Tech Yes City is a total joke, and this is far from the first time he's tried to stir poo poo about a subject he barely understands. A lot of the C-tier techtubers are borderline tech-illiterate, the RedGamingTechs and MindBlankTech and UFD Tech etc, and especially when they come up with some “actually AMD is way better than NVIDIA for [counterintuitive or bizarre reason]” it’s usually best to just take them with a massive grain of salt or just flatly ignore them, because they’re usually straining to reach some predetermined conclusions. I remember a couple years ago it was “nvidia isn’t rendering as sharp/worse colors than AMD!!!” because someone physically pointed a camera at the screen and it wasn’t autofocusing correctly, and it started a whole thing that took 6-12 months to dissipate. Or that AMD has better latency (no, it’s worse). Etc.
|
# ? Jun 29, 2023 14:02 |
|
well, thanks for the responses. I'm at least pleased that my skepticism twigged on that sort of thing.
|
# ? Jun 29, 2023 14:08 |
|
Whats the deal with these new Xeon W chips? Not many reviews and they seem to be the new HEDT chips. Was poking around at building a home lab and theyre intriguing but not sure how much more when you can get a 5950X on sale or even just using the 12900Ks. No E cores i suppose is useful here.
|
# ? Jul 2, 2023 18:23 |
|
doomisland posted:Whats the deal with these new Xeon W chips? Not many reviews and they seem to be the new HEDT chips. Was poking around at building a home lab and theyre intriguing but not sure how much more when you can get a 5950X on sale or even just using the 12900Ks. No E cores i suppose is useful here. The deal is 112 pcie5 lanes
|
# ? Jul 2, 2023 18:33 |
|
Apple should have just updated the MP to use these instead of the burning dumpster fire that is the ASi Mac Pro.
|
# ? Jul 2, 2023 19:51 |
|
SourKraut posted:Apple should have just updated the MP to use these instead of the burning dumpster fire that is the ASi Mac Pro. They were eager to move onto in-house solutions for PR, but the bet on an M2 Extreme / four M2 Max chips all working in tandem fell apart due to manufacturing problems. Thus, the Mac Pro 2023 is the fallback: literally an Ultra Studio with PCIe 4 support enabled by switching, no external GPU support, and 64GB of non-upgradable RAM starting at seven thousand dollars. I know you know how insane this is, but I’m still aghast.
|
# ? Jul 2, 2023 20:44 |
|
moving their final product line over to ARM, even if it's half assed, also started the countdown to them being able to retire intel support altogether releasing a new xeon mac pro would have kicked that down the road by years
|
# ? Jul 2, 2023 21:00 |
|
It's € 8.399 here. But it's Apple, they cornered the "price insensitive buyers" market and will probably sell enough of those at gently caress you margins to make up for the cost.
|
# ? Jul 2, 2023 21:57 |
|
even for price insensitive buyers it's a weird proposition though, the specs are the same as the mac studio that's $3000 cheaper, so all you're getting for the extra cost is pcie slots it doesn't support GPUs, so the only reason you'd want pcie slots is because pcie devices tend to be cheaper than their thunderbolt counterparts, but is anyone going to save enough that way to offset $3000?
|
# ? Jul 2, 2023 22:01 |
|
Hasturtium posted:They were eager to move onto in-house solutions for PR, but the bet on an M2 Extreme / four M2 Max chips all working in tandem fell apart due to manufacturing problems. Thus, the Mac Pro 2023 is the fallback: literally an Ultra Studio with PCIe 4 support enabled by switching, no external GPU support, and 64GB of non-upgradable RAM starting at seven thousand dollars. I know you know how insane this is, but I’m still aghast. Yeah, it's horrible. It almost feels like an excuse for them to point to poor sales in a year or two in order to end the Mac Pro line anyway. repiv posted:moving their final product line over to ARM, even if it's half assed, also started the countdown to them being able to retire intel support altogether At this point, they still have a substantial number of Intel-based Macs out there, and they developed Rosetta 2 themselves (plus the hardware implementation to support it), so it's less urgent to do so. Of course Apple could surprise us, but it seems like they'd be more than happy to just force people away via software feature segmentation, as we're already seeing with Sonoma. They seem more interested with dropping support for anything that doesn't have the T2 chip. I would assume though that the last of the Intel Macs will see support until roughly 2026 or 27 OS-wise, probably the 2019 Mac Pro, and then security updates after that.
|
# ? Jul 2, 2023 22:08 |
hobbesmaster posted:The deal is 112 pcie5 lanes
|
|
# ? Jul 3, 2023 11:43 |
|
intel updated their architecture guide and it looks like client arrow lake still isn't going to have AVX512 support tired: an intel extension having slow adoption because AMD doesn't support it yet wired: an intel extension having slow adoption because intel doesn't support it yet
|
# ? Jul 3, 2023 18:00 |
|
Seriously, though, how much of Intel not having AVX512 support because of AMD? I get that there are shifted dynamics in the server space, but AMD only commands about 20% of the server market, and I'm sure *somebody* out there still religiously adheres to "nobody ever got fired for buying SwissArmyDruid fucked around with this message at 18:32 on Jul 3, 2023 |
# ? Jul 3, 2023 18:09 |
|
BlankSystemDaemon posted:Needs more PCIe lanes. Never enough PCIe lanes!! Also my current dream is some kind of EDSFF x16 connector on steroids that has power delivery pins so you can have MCIO from the motherboard to a backplane you can slot your various boards in like GPU, NICs etc.. I'm just real sick of CEM connectors!
|
# ? Jul 3, 2023 18:53 |
The complex history of the Intel i960 RISC processor is a nice little read for a branch not taken. It's also interesting that its Ada properties are shared by the Rational R1000, a machine which played a central role in computing in Denmark (and which is currently in the process of being restored). BlankSystemDaemon fucked around with this message at 19:12 on Jul 3, 2023 |
|
# ? Jul 3, 2023 19:08 |
|
|
# ? Apr 28, 2024 06:44 |
|
SwissArmyDruid posted:Seriously, though, how much of Intel not having AVX512 support because of AMD? I get that there are shifted dynamics in the server space, but AMD only commands about 20% of the server market, and I'm sure *somebody* out there still religiously adheres to "nobody ever got fired for buying Isn’t that as much about supply limitations?
|
# ? Jul 3, 2023 19:08 |