|
From the GPU thread. This matters far more to me than raw fps. Just one benchmark tho, hopefully people do more and it's not just in 1 game.
|
#
?
Mar 6, 2017 11:26
|
|
|
|
| # ? May 21, 2024 21:29 |
|
|
FaustianQ posted:Isn't there a disadvantage to doing this though? Every idiotic piece of software with per-CPU licensing or restrictions, including win10 home.
|
|
#
?
Mar 6, 2017 11:33
|
|
|
FaustianQ posted:Isn't there a disadvantage to doing this though? To this point, AMD has failed to demonstrate the value of the two-core-complex design, and with that said, probably should have presented it as two processors, instead of having the problem they're facing right now, with the OS thinking that the L3 is monolithic when it actually isn't. edit: To clarify: Windows OSes already have scheduler programming that recognizes how to handle juggling resources between two sockets on a motherboard. Rather than coming out with the current situation, and then having to wait for a hotfix to the scheduler so that it doesn't try to treat the L3 as contiguous when it isn't, maybe AMD should have just had motherboard partners present the 8-core parts as two 4-core parts. AMD has not been forthcoming with an explanation why they didn't do this in the first place. (I'm sure the endgame is probably to use the core complex as a building block for 4-core parts and APUs down the line, but right now, this 4+4 layout needs to be worked around before it can be declared as "good". Right now, the best that Ryzen has going for it is "interesting".) SwissArmyDruid fucked around with this message at 12:10 on Mar 6, 2017 |
|
#
?
Mar 6, 2017 11:44
|
|
|
Is it done to make fusing off core groups easier to achieve? E- partly GRINDCORE MEGGIDO fucked around with this message at 12:16 on Mar 6, 2017 |
|
#
?
Mar 6, 2017 12:14
|
|
|
One of the comments on that Tech City video:quote:The program that Wendell is talking about for fixing process thread affinity (override WIn 10) already exists. It's called Process Lasso. And yes, folks on tech forums already tested it, MASSIVE performance gains for Ryzen when Win 10 cannot stuff up CCX1 vs CCX2 thread movement. Sounds interesting, but then they didn't link to anything so I have no idea what the numbers are.
|
|
#
?
Mar 6, 2017 12:18
|
|
|
Anime Schoolgirl posted:This was the big goal of the Zen core (as well as have a new architecture for servers since 2010, yeesh). If they can translate that to ultrabook-usable dual core APUs (and the binned Polaris 11 parts are comically efficient), they're golden in mobile. Yeah apparently Ryzen absolutely smashes the i7-6700HQ in singlethreaded while running less hot and hungry. Raven Ridge looks to be absolutely killer, and if they have a good mobile and good server CPU they can afford to be a bit behind in desktop until later. Truga posted:
I want to see how this pans out in R5 and R3 CPUs, because sure a 6600 might hit 100+ frames but if an R3 1200X gets better frametimes for 150$, why bother with the 6600K in a budget system? If the R5s are smoother for the same price, why bother with the i5? I mean, keep in mind this is with the scheduler treating the L3 as monolithic when it isn't and going spastic about thread assignment, so I don't think it's related to having double the L3.
|
|
#
?
Mar 6, 2017 13:19
|
|
|
FaustianQ posted:Also on Anandtech forums, something interesting was pointed out by The Stilt Put...Put an underclocked 1800X in laptop chassis? My 5700HQ is 47W... I would like to do science with this  Do you have the link for me to dumbly ogle? E: found it NewFatMike fucked around with this message at 14:01 on Mar 6, 2017 |
|
#
?
Mar 6, 2017 13:56
|
|
|
Maxwell Adams posted:Does Linux have these thread scheduling problems? I want to see some Linux gaming benchmarks. I want someone to establish the high-end processor of choice for Linux gaming. The Linux kernel is reported to already have all the major fundamentals patched. Pablo Bluth fucked around with this message at 14:07 on Mar 6, 2017 |
|
#
?
Mar 6, 2017 13:58
|
|
|
FaustianQ posted:I want to see how this pans out in R5 and R3 CPUs, because sure a 6600 might hit 100+ frames but if an R3 1200X gets better frametimes for 150$, why bother with the 6600K in a budget system? If the R5s are smoother for the same price, why bother with the i5? Yeah, same. If R5 in gaming are similar to R7 (little reason to expect otherwise right now) and retain similar clocks, it'd be the best thing to happen to budget pcs since K7.
|
|
#
?
Mar 6, 2017 14:18
|
|
|
Our favourite Scot talks about the Ryzen launch: https://www.youtube.com/watch?v=ylvdSnEbL50
|
|
#
?
Mar 6, 2017 15:41
|
|
|
Truga posted:
That's just the effect of more (fast) cores in that specific game. This does show that we need more affordable 8 core cpus for gaming I agree. This is the 'smoothness' some reviewers were talking about.   Another example  With more cores 
|
|
#
?
Mar 6, 2017 16:00
|
|
|
SwissArmyDruid posted:
Or you know, MS had to work on the scheduler...and AMD launched too loving early so that they didn't miss Q1 (a stupid decision in the end I think). I don't see anything wrong with the actual architecture, and the performance shows it (after all). Also, the endgame for this part I think is a Xeon perf/$/w competitor for some parts of the server market (the 16/24/32c Naples chip). The APUs will need vega to be efficient if they want to succeed. Dante80 fucked around with this message at 16:13 on Mar 6, 2017 |
|
#
?
Mar 6, 2017 16:03
|
|
|
Dante80 posted:The APUs will need vega to be efficient if they want to succeed. Do they really? I think some vendors would be happy with a 4/6-core Zen + Polaris APU. Vega would be the icing in the cake. Preliminary low wattage testing for Zen looks excellent. I can totally see a 65W Zen + Polaris chip in one of the next MBPs. Vega + shared HBM would be the icing on the cake. That would save a lot of space (separate dGPU + VRAM chips) on the PCB that could be used for a bigger battery.
|
|
#
?
Mar 6, 2017 16:21
|
|
|
I'd like badass APUs too but battery sizes/capacities are more constrained by laptop thickness and footprint more than motherboard size these days.
|
|
#
?
Mar 6, 2017 16:31
|
|
|
Seamonster posted:I'd like badass APUs too but battery sizes/capacities are more constrained by laptop thickness and footprint more than motherboard size these days. Lets make laptop thickness great again? Seriously though, gently caress Apple and gently caress anyone selling "pro" laptops that put aesthetics ahead of power.
|
|
#
?
Mar 6, 2017 16:39
|
|
|
I mean the XPS 15 has a 97Wh battery in a small form factor with a dGPU, so hopefully more OEMs jump on that because that with a badass APU is pretty much all I want for Christmas.
|
|
#
?
Mar 6, 2017 16:45
|
|
|
You can still get pro laptops that aren't poo poo, though they do cost just as much as they did 5 years ago while everything else comparable went down by 50%, it kinda sucks. I wish the new vaio Z didn't blow, old ones were literally 13" workstations.
|
|
#
?
Mar 6, 2017 16:46
|
|
|
Anime Schoolgirl posted:This was the big goal of the Zen core (as well as have a new architecture for servers since 2010, yeesh). If they can translate that to ultrabook-usable dual core APUs (and the binned Polaris 11 parts are comically efficient), they're golden in mobile. Ultrabook quads you mean.
|
|
#
?
Mar 6, 2017 16:46
|
|
|
WCCFtech rumor: Add-In-Board partners ramp production as global AM4 motherboard supply buckles under strong Ryzen demand
|
|
#
?
Mar 6, 2017 17:00
|
|
|
NewFatMike posted:I mean the XPS 15 has a 97Wh battery in a small form factor with a dGPU, so hopefully more OEMs jump on that because that with a badass APU is pretty much all I want for Christmas. Why do you particularly care how good the iGPU is if then system has a dGPU?
|
|
#
?
Mar 6, 2017 17:01
|
|
|
It was really more that they're able to fit such a large battery while having space taken up by a discrete GPU than APU+GPU comment, I definitely could have been more clear on that
|
|
#
?
Mar 6, 2017 17:06
|
|
|
Platystemon posted:Why do you particularly care how good the iGPU is if then system has a dGPU? so you can dump the dgpu
|
|
#
?
Mar 6, 2017 17:07
|
|
|
Measly Twerp posted:Lets make laptop thickness great again? This made me think of a recent article about the Iris Skylake (6560U, I think) model of the XPS 13 that talked about how under combined GPU+CPU load, the core temperature goes up to 90+C and throttling eats most of the potential performance gains from Iris. The author saw great improvement from repasting the processor and especially from adding thermal pads to connect heatpipes with the aluminum chassis. Upon reading this I thought "drat, I should check on temperatures for my Latitude E7270 - it's a similar processor and a similar form factor!" I found though that the slightly increased thickness of the Latitude hides a far superior cooling solution and even with Furmark + Intel XTU's CPU test, I only get up to around 65C. Between that and being able to upgrade the RAM, I continue to be glad that I went for the business model. Eletriarnation fucked around with this message at 17:15 on Mar 6, 2017 |
|
#
?
Mar 6, 2017 17:12
|
|
|
Measly Twerp posted:Our favourite Scot talks about the Ryzen launch: That's a interesting take, the stuff about the FX is really eye-opening. Demolishes some long-held assumptions about gaming benchmarking. I don't know that I agree with his statement that games are heavily optimized for intel CPUs. If anything, any cross-platform game (or game built on a cross-platform engine) would have the most optimization work done for Jaguar CPUs, since that's the platform that really needs it. But mostly I think that the bulk of optimization is graphics related and CPUs are left to fend for themselves.
|
|
#
?
Mar 6, 2017 17:17
|
|
|
Rastor posted:WCCFtech rumor: Canada has basically no boards, very few were available at launch. You can walk in to any store to grab the CPU today. The best chance of building a box today is driving across the border to something like a Detroit Microcenter. Great launch :iamafag:
|
|
#
?
Mar 6, 2017 17:23
|
|
|
Rastor posted:WCCFtech rumor: It should read AMD had enough initial product meet demand, surprising AIB. Really though, you'll see AIB care greatly when thos design wins come in. incoherent fucked around with this message at 21:08 on Mar 6, 2017 |
|
#
?
Mar 6, 2017 17:28
|
|
|
Measly Twerp posted:Our favourite Scot talks about the Ryzen launch:  This overturns my worldview.
|
|
#
?
Mar 6, 2017 17:35
|
|
|
Klyith posted:That's a interesting take, the stuff about the FX is really eye-opening. Demolishes some long-held assumptions about gaming benchmarking.
|
|
#
?
Mar 6, 2017 17:40
|
|
|
Alereon posted:It seems like he's basically observing that graphics drivers got better at multi-threading to alleviate CPU bottlenecks over a few years. None of this is new information, but some people get really hung up on the percentages, as if a game is 10-15% less fun on a Ryzen than a 7700k.
|
|
#
?
Mar 6, 2017 18:03
|
|
|
Okay yeah bad benchmarks exist but if anything FX-series CPUs were worse for gaming than those benchmarks indicated, so he is making a kind of stupid point.
|
|
#
?
Mar 6, 2017 18:43
|
|
|
Alereon posted:Okay yeah bad benchmarks exist but if anything FX-series CPUs were worse for gaming than those benchmarks indicated, so he is making a kind of stupid point. He addressed that starting at about 14:29 though.
|
|
#
?
Mar 6, 2017 18:48
|
|
|
Alereon posted:Okay yeah bad benchmarks exist but if anything FX-series CPUs were worse for gaming than those benchmarks indicated, so he is making a kind of stupid point.
|
|
#
?
Mar 6, 2017 18:52
|
|
|
Rastor posted:He addressed that starting at about 14:29 though.
|
|
#
?
Mar 6, 2017 18:58
|
|
|
As near as I can tell there were two points in that video: 1: Low resolution tests of current games aren't a good way to anticipate future performance, because future performance will likely take better advantage of more cores. I think this is a decent point, but there should be SOME games today that take advantage of multiple cores reasonably well and... the 7700K still beats the 1800X in those (with 6900K on top, so the games DO use those cores.) 2: This advantage Intel chips have over Zen in gaming is the reverse of other workloads so it's all optimization~! I mean, this could be true, but its just stated with no evidence at all. That said, it shouldn't take long for the bug fix / optimization evidence to start rolling in, even if it's only for some games, and only for DX12/Vulkan titles. Personally, I think the best argument for more cores is one he only kinda alludes to. We're in a dead-zone on core scaling right now. We've just got DX12/Vulkan, and we've got a few games that use these APIs well, but we don't have any games or game engines that are really designed to use lots of core for non-rendering work. So today we're in a partial state of "well, more cores helps a little (finally!), but we're not really making full use of them... yet." That probably won't change until we get the next generation of consoles, and game engines designed for them.
|
|
#
?
Mar 6, 2017 19:36
|
|
|
crazypenguin posted:So today we're in a partial state of "well, more cores helps a little (finally!), but we're not really making full use of them... yet." That probably won't change until we get the next generation of consoles, and game engines designed for them. I wonder if the 4-core to 4-core comparisons, whenever AMD releases those, will show a more accurate comparison given that's essentially what the current consoles have.
|
|
#
?
Mar 6, 2017 19:41
|
|
|
crazypenguin posted:As near as I can tell there were two points in that video: It is really a self-defeating argument I think, especially if you place it due to AMD loyalty (I'm not saying that guy is doing it due to that, but a lot of others do). I mean, think about this. AMD is going to roll out the CPUs that they intend to sell to gamers on a perf/$ basis, via the R5 launch (1600X 6c and 1500X 4c). Those are salvaged chips, intended to target i5 (and top i3) prices respectively. If AMD manages to get the same gaming IPC performance out of those, they will be more or less identical to R7 in that particular market (since current games don't scale well with that many threads), while costing a lot less. Also, they may get a much better launch this time, due to the platform already being live for a month or two (and some OS and BIOS optimization happening). What are the AMD fans going to do then...snob them because they are not "future proof"? Dante80 fucked around with this message at 19:49 on Mar 6, 2017 |
|
#
?
Mar 6, 2017 19:47
|
|
|
Those 1080p tests are extremely relevant, right now, to people who buy high refresh rate monitors. There are a bunch of 144 hz panels out there, and most of them are only 1080p. Once you put down the cash for one of those monitors, you look at 100 fps as a pretty good start.
|
|
#
?
Mar 6, 2017 19:47
|
|
|
Going high FPS is the only "exotic" display tech Blizzard allows in their games, too
|
|
#
?
Mar 6, 2017 19:53
|
|
|
Has any goon here got a Ryzen build yet?
|
|
#
?
Mar 6, 2017 20:08
|
|
|
|
| # ? May 21, 2024 21:29 |
|
|
My 'preorder' cpu ships March 8th, my 'preorder' motherboard ships March 13th 
|
|
#
?
Mar 6, 2017 20:10
|
|