|
nm. i can't read, should go to sleep.
Rinkles fucked around with this message at 07:14 on Nov 15, 2021 |
# ? Nov 15, 2021 07:10 |
|
|
# ? Jun 9, 2024 00:39 |
|
For anyone with a Coffee Lake processor and a case of upgraditis, this is my 9900KF (4.9GHz all core) versus my new 12700K on stock. The rest of the system is a 3080 FE, C14/3600 DDR4 and an NVME drive. By far the most noticeable improvement has been Windows chugging less immediately after boot or when quickly alt-tabbing between a game and other programs.
|
# ? Nov 15, 2021 22:08 |
|
Would you take a 10700k over a 11600k? Gamer's Nexus had the 11600k ahead in a lot of their gaming benchmarks, and there's the PCIe 4 support (I'm unclear how relevant it'll end up being. DirectStorage will support PCIe 3, iinm). But the 10700k is an 8 core, and probably more OCable. They're sometimes in a similar price range. Rinkles fucked around with this message at 23:29 on Nov 15, 2021 |
# ? Nov 15, 2021 23:27 |
|
If you like gaming, I would consider a 12600k + DDR4 board (the MSI Pro Z690-A that Buildzoid likes, for example) if your budget is in that range. Just goofing with current PC Part Picker prices, the board+CPU prices with the cheapest mobo I would consider are: $356 - 11600k $428 - 10700k $509 - 12600k $150 increase isn't that huge by raw price, especially diluted as part of a total system (if you are doing that) and with the knowledge that the 12600k will absolutely have a longer usable lifetime when you are starting from today. More cores + better cores + more modern platform + whats bound to be a more popular platform + better in socket upgrade options in the future with the news items about Raptor Lake also supporting DDR4 all make a good argument for the price increase, to me.
|
# ? Nov 16, 2021 08:11 |
|
https://twitter.com/momomo_us/status/1460244278647480324 https://twitter.com/VideoCardz/status/1460255159502704643 and the other is obviously Raptor Lake
|
# ? Nov 16, 2021 08:48 |
|
Gotta say, Fishhawk Falls is the weirdest codename Intel has had for a while Kinda hyped to see how it turns out. Are the rumours pointing to all big cores (golden cove)?
|
# ? Nov 16, 2021 17:05 |
|
Cygni posted:If you like gaming, I would consider a 12600k + DDR4 board (the MSI Pro Z690-A that Buildzoid likes, for example) if your budget is in that range. Just goofing with current PC Part Picker prices, the board+CPU prices with the cheapest mobo I would consider are: No I agree, but an extra $100 would already be pushing it, and the new motherboards are expensive (more than you outlined because I need wifi) and much less likely to get Black Friday sales than the LGA1200 mobos. Idk, maybe if I'm in a good mood I'll splurge, but as we talked about on the last page, for gaming the 12000k isn't likely to make a big difference, especially without the budget for a high end GPU.
|
# ? Nov 16, 2021 17:21 |
|
FYI, I managed to get Best Buy to price match the $200 11600K Micro center offer. It took two times. I didn't want to waste CS' time, but the first guy seemed to not understand the Micro center page (he thought the part was out of stock). But your mileage may definitely vary. And I doubt the Tustin specific sale ($150 10700K) would fly because they asked for my zip code multiple times.
|
# ? Nov 16, 2021 22:28 |
|
Some cool pictures of upcoming products. Meteor Lake, Ponte Vecchio, Sapphire Rapids etc: https://www.cnet.com/pictures/a-look-inside-intels-mammoth-arizona-chipmaking-fab/ Guys i think this chiplet thing may just be the future afterall
|
# ? Nov 18, 2021 21:54 |
|
Looking forward to people who are locked into the "big pea" mindset killing those chips. Or rather, the desktop and HEDT versions that are their descendants. (yes, I know the gold ones are meant for supercomputers)
|
# ? Nov 18, 2021 22:05 |
|
DPPH
|
# ? Nov 18, 2021 22:25 |
|
Is it new that Intel is allowing reporters to take so many pictures of unreleased chips and fab equipment? Although it does like it has been heavily curated, they seem to have carefully omitted any pictures of the steppers.
|
# ? Nov 18, 2021 22:28 |
|
I thought chiplets were more delicate. I guess if the balls are tens of microns apart there's more leeway to treat them under less than cleanroom conditions. I thought they were all assembled pretty much after being cut from the wafers, still inside the originating fab. e: btw tech jesus did benchmark the 12900k IGP. There's as good as no difference between them, wtf intel?
|
# ? Nov 18, 2021 22:45 |
|
karoshi posted:e: btw tech jesus did benchmark the 12900k IGP. There's as good as no difference between them, wtf intel? Intel no longer has apple to ride their rear end for bad quality onboard.
|
# ? Nov 19, 2021 00:01 |
|
Beef posted:Is it new that Intel is allowing reporters to take so many pictures of unreleased chips and fab equipment? Although it does like it has been heavily curated, they seem to have carefully omitted any pictures of the steppers. It is absolutely a change for them, supposedly part of Gelsinger's new direction for the company. It also tends to be what tech companies that are behind their competitors do to stir up positive press and investor confidence. I'm sure thats not related, no no. karoshi posted:e: btw tech jesus did benchmark the 12900k IGP. There's as good as no difference between them, wtf intel? It's the same 32EU Xe design as Rocket Lake. The parts with the big 96EU designs are mobile only. You don't want to use any of the integrated graphics options on desktop, AMD or Intel, for any gaming thats remotely modern, honestly.
|
# ? Nov 19, 2021 00:41 |
|
Cygni posted:It's the same 32EU Xe design as Rocket Lake. The parts with the big 96EU designs are mobile only. You don't want to use any of the integrated graphics options on desktop, AMD or Intel, for any gaming thats remotely modern, honestly. Well, congratulations to Intel on delivering an IGP that's not limited by RAM bandwidth.
|
# ? Nov 19, 2021 01:30 |
|
I was looking at Hardware Unboxed and Gamer's Nexus benchmarks of the 11600K, and noticed there's a pretty big disparity between the results for the two games that overlapped in their testing. HU has the 11600K matching the 10600K in F1 2020, whilst the 11-chip is fairly significantly ahead for GN And in Siege, GN's numbers are again favorable for the 11600K, where it's almost as strong as the 5600X, whilst in HU's testing the 5600X is way ahead There's a similar pattern in the application benchmarks. Each YouTube channel uses different pc components in their testing rigs, and different graphical presets for games, but I was surprised at just how consistently the 11600K looked better in GN's benchmarks. Could it be silicon lottery, or more down to differences in testing?
|
# ? Nov 19, 2021 02:31 |
|
Rinkles posted:Each YouTube channel uses different pc components in their testing rigs, and different graphical presets for games, but I was surprised at just how consistently the 11600K looked better in GN's benchmarks. Could it be silicon lottery, or more down to differences in testing? Most of the above, but I doubt silicon lottery makes a significant difference at stock speed. Different game settings used for testing probably make up a lot of it. Helter Skelter fucked around with this message at 04:59 on Nov 19, 2021 |
# ? Nov 19, 2021 04:56 |
|
Rinkles posted:I was looking at Hardware Unboxed and Gamer's Nexus benchmarks of the 11600K, and noticed there's a pretty big disparity between the results for the two games that overlapped in their testing. Probably ram, GN uses the same kit of 3200c14 BDie that they manually tune the subtimings on. Ram subtimings can have a massive impact in some games, I got 10%+ in FC6 going from 4000c17 XMP to 3900c14 with tuned subs.
|
# ? Nov 19, 2021 05:54 |
|
They are building new highly water-intensive fabs... in Arizona????
|
# ? Nov 19, 2021 09:22 |
|
it's my understanding that Arizona is rather seismologically stable, and consequently already has a bunch of existing IT corps and data centers operating in the state, as points in the Pro column, even if being in the middle of the drat desert might otherwise make it sound like a really bad idea
|
# ? Nov 19, 2021 09:27 |
|
gradenko_2000 posted:it's my understanding that Arizona is rather seismologically stable, and consequently already has a bunch of existing IT corps and data centers operating in the state, as points in the Pro column, even if being in the middle of the drat desert might otherwise make it sound like a really bad idea
|
# ? Nov 19, 2021 09:54 |
|
There is a fuckton of water purification infrastructure at fab sites by necessity. Intel is already flowing back 95% or something of the water intake and they are aiming for 100+% Edit: https://www.cnbc.com/amp/2021/06/04/why-intel-tsmc-are-building-water-dependent-chip-plants-in-arizona.html Beef fucked around with this message at 11:27 on Nov 19, 2021 |
# ? Nov 19, 2021 11:24 |
|
Plus Intel already has fabs and empty fab space in the PHX area. Also, (and I cannot explain why) water was pretty loving cheap when I lived in PHX for some reason.
|
# ? Nov 19, 2021 16:54 |
|
gradenko_2000 posted:it's my understanding that Arizona is rather seismologically stable, and consequently already has a bunch of existing IT corps and data centers operating in the state, as points in the Pro column, even if being in the middle of the drat desert might otherwise make it sound like a really bad idea water is like skilled talent you want to trap as much of that as possible
|
# ? Nov 19, 2021 17:23 |
|
Intel currently does on-site treatment for their waste stream, but they still send quite a bit of effluent out to Chandler. The water that they can re-purpose from their on-site treatment systems, typically gets used for cooling purposes, and they purchase potable water from Chandler and then do ultra purification on it for manufacturing use. TSMC will eventually do some on-site treatment and re-use, ultimately for cooling purposes also, but that is a few years out from when they begin manufacturing, and I think until then, most of the waste will go to Phoenix. They do want to do ultra purification also, but that'll be even later, since they're still working on the ultra purification reuse approach in Taiwan and are using that as a pilot of sorts.
|
# ? Nov 19, 2021 20:50 |
|
I was gonna head down to Microcenter today to get one of those insane $150 10700Ks for a friend, but i regret to announce that they are sold out and the price is back up to $220. They managed to last a pretty long time at that price so that must have been a hell of a shipment they had.
|
# ? Nov 20, 2021 21:30 |
|
12600K for $270 at Microcenter My MC was already low on stock the last few days, so there's no way there'd be any left now. Worth noting it's out of stock in Best Buy and Amazon, so maybe it's not selling that poorly after all. Also Z690 prices haven't budged. Rinkles fucked around with this message at 13:22 on Nov 23, 2021 |
# ? Nov 23, 2021 13:03 |
|
Making the jump to the 12900k and going with DDR4. Unfortunately my current 9700k build has some garbage 3000CL15 RAM i bought back in 2018 when the memory market was pretty hosed pricewise, so I had to buy a new kit anyway. People are saying that the MSI boards are pretty solid for DDR4 so I'm going with the Pro Z690-A, also helps that it's quite affordable.
|
# ? Nov 24, 2021 13:21 |
|
You aren’t gonna believe this, but the 2nd wave of DDR5 modules look much much better than the launch options. Whomst could have predicted! https://youtu.be/I5pXSBus8pQ
|
# ? Nov 26, 2021 21:19 |
|
Interesting that it's 1.25v compared to 1.35v on existing DDR4 sticks
|
# ? Nov 26, 2021 22:02 |
|
Zedsdeadbaby posted:Interesting that it's 1.25v compared to 1.35v on existing DDR4 sticks DDR5 was designed to use lower voltages, no?
|
# ? Nov 26, 2021 22:04 |
|
Rinkles posted:DDR5 was designed to use lower voltages, no? 1.1v-1.6v, so the range is quite large. ADATA or whatever who did the DDR5-12600 announcement a while back said they were 8400 1.1v sticks running at 12600 1.6v. Future is going to go insane for DDR5. In other stupid news, I'm water-cooling my bdie to see if I can hit 30ns latency.
|
# ? Nov 27, 2021 00:38 |
|
Seems H670 and B660 boards (and the lower end ADL parts to go with them) are fairly imminent, as the lineups and full prices are leaking. In related news, Intel and AMD both have presentations at CES on January 4th. https://twitter.com/g01d3nm4ng0/status/1466812876585598976
|
# ? Dec 3, 2021 21:31 |
|
I replacing the motherboard I installed just last week. Would it be a bad idea to keep the thermal paste? Just asking since it's still fresh and there's quite a bit of it. If I'm applying a new layer, can I use lens wipes to clean the surfaces?
|
# ? Dec 4, 2021 00:38 |
|
Rinkles posted:I replacing the motherboard I installed just last week. Would it be a bad idea to keep the thermal paste? Just asking since it's still fresh and there's quite a bit of it. Microfiber cloth and 70% isopropyl is what I use - you can't be sure if those lens cleaning cloths won't leave behind strands.
|
# ? Dec 4, 2021 01:07 |
|
Coffee filters work well for cleaning thermal paste. I would reapply if you have more paste lying around.
|
# ? Dec 4, 2021 03:25 |
|
Yeah, most of the point of reapplying thermal paste any time you expose it is making sure you don't have any voids/bubbles in the seal. Even if the paste itself is fresh, sticking two surfaces back together with paste that is already spread out and then pulled apart is more likely to trap bubbles inside than spreading out from a central blob/X/stripe. Not sure how much difference it makes in practice especially if your heatsink is overkill, but I'd also reapply if I had the paste on hand.
|
# ? Dec 4, 2021 03:37 |
|
Cygni posted:Seems H670 and B660 boards (and the lower end ADL parts to go with them) are fairly imminent, as the lineups and full prices are leaking. In related news, Intel and AMD both have presentations at CES on January 4th. do we know if the H and B boards will support memory tuning like the 500 series did? Rinkles posted:I replacing the motherboard I installed just last week. Would it be a bad idea to keep the thermal paste? Just asking since it's still fresh and there's quite a bit of it. nthing that you should generally reapply paste whenever you pull the cooler apart from the CPU no matter how "fresh" it might still look
|
# ? Dec 4, 2021 03:47 |
|
|
# ? Jun 9, 2024 00:39 |
|
I hosed up and ended up repasting three times, because I noticed that the cooler was loose, but only by the second attempt did I realize it was because I forgot to screw in the mounting brackets. At least this was before I moved on to do anything else. A lot of wasted paste, though by the end I was more confident about how much (or little) was actually needed.
|
# ? Dec 4, 2021 05:35 |