|
PC LOAD LETTER posted:Based on the information we have right now (which admittedly isn't much) his post isn't all that bad to me. If AMD pulls off those levels of performance with those prices and power usage they should do alright. Unless nV messes up somehow, either on pricing or performance, I don't see AMD clawing back huge amounts of dGPU marketshare though.
|
#
?
Jan 31, 2016 21:57
|
|
|
|
| # ? May 30, 2024 23:59 |
|
Subjunctive posted:But if it's AIO, how are you supposed to bleed bubbles out? This can't be the desired state of operation. You don't, the AIO depends on orientation to deal with the small amount of air inside. If you put the rad/rez combo up above the pump/block/card the bubbles naturally travel up and don't mess with the pump and block, most people put the rad in the back of the case as an exhaust since that position is naturally above the card in most cases.
|
|
|
#
?
Jan 31, 2016 22:01
|
|
|
AVeryLargeRadish posted:You don't, the AIO depends on orientation to deal with the small amount of air inside. If you put the rad/rez combo up above the pump/block/card the bubbles naturally travel up and don't mess with the pump and block, most people put the rad in the back of the case as an exhaust since that position is naturally above the card in most cases. Yeah, but I have 2 cards, sigh. I can't believe I'm considering a custom loop now.
|
|
#
?
Jan 31, 2016 23:03
|
|
|
Subjunctive posted:Yeah, but I have 2 cards, sigh. I can't believe I'm considering a custom loop now. Tell EVGA that its running at 90 degrees at 100% fan speed. Its broken, and they will send you another. Orientation can matter but this is just broken. A ton of people have radiators mounted on top (I have too, specifically for the GPU). The performance difference can be there but this isn't that and its really a much bigger deal for CPU's anyway.
|
|
#
?
Jan 31, 2016 23:31
|
|
Subjunctive posted:Yeah, but I have 2 cards, sigh. I can't believe I'm considering a custom loop now. Unfortunately a custom loop is one of the few truly quiet solutions for multi-card setups. Just so you know you will be looking at spending $450-$550 for a dual card loop.
|
|
|
#
?
Feb 1, 2016 00:05
|
|
|
AVeryLargeRadish posted:Unfortunately a custom loop is one of the few truly quiet solutions for multi-card setups. Just so you know you will be looking at spending $450-$550 for a dual card loop. More than that, GPU blocks alone will be $150 each, add $50-100 for cpu block, say $100 -200 for rads, 60-120 for fans, $50-$100 for fittings, $150 for pump + res. poo poo adds up fast.
|
|
#
?
Feb 1, 2016 00:55
|
|
Don Lapre posted:More than that, GPU blocks alone will be $150 each, add $50-100 for cpu block, say $100 -200 for rads, 60-120 for fans, $50-$100 for fittings, $150 for pump + res. Actually I calculated that based on the actual costs, I was also just including costs for a dual-GPU loop, not the CPU too. The GPU blocks can be had for $125 each, a single 240mm rad will handle a couple GPUs and cost $60, two fans at $20 each is $40, ten decent compression fittings at ~$5 each is $50 and a combined pump and res can be had for $100, flexible tubing is cheap so even if I include some extra overhead for some PT Nuke and maybe a couple fuckups along the way it comes to like $550.
|
|
|
#
?
Feb 1, 2016 01:24
|
|
|
Followup to my "second monitor looks like rear end when Catalyst drivers are installed on my XFX R9 290, regardless of cable or output type" story: I recently acquired, for a very low price, an ASUS DirectCUII R9 290 that was one of the early runs of perfectly-fine 290X chips softlocked to 290 specs. I think I mentioned this earlier in the thread. Well, the previous owner had unlocked it to a 290X. Today I rearranged my GPUs so that my XFX 290 would be the secondary and this 290X would be the primary, thinking this would be a good way to get a little more oomph in games that don't handle CrossFire well. Lo and behold, my monitor that Subjunctive said would only work nicely if my setup was cooled with holy water looks perfectly fine plugged into the ASUS 290X. I don't get it.
|
|
#
?
Feb 1, 2016 06:20
|
|
|
Subjunctive posted:But if it's AIO, how are you supposed to bleed bubbles out? This can't be the desired state of operation. Generally speaking, you don't. You cross your fingers and hope that they stay in that part of the radiator where the hoses go into, and that they don't get stuck in the rest of the radiator or pumpblock. Some AIOs have service ports? (I think I'm thinking of the Raijintek one does, the one you can self-dye? And the EKWB Predator, obviously.) But by and large, let alone where GPU AIOs are concerned, no.
|
|
#
?
Feb 1, 2016 07:33
|
|
|
Kazinsal posted:Followup to my "second monitor looks like rear end when Catalyst drivers are installed on my XFX R9 290, regardless of cable or output type" story: I recently acquired, for a very low price, an ASUS DirectCUII R9 290 that was one of the early runs of perfectly-fine 290X chips softlocked to 290 specs. I think I mentioned this earlier in the thread. Well, the previous owner had unlocked it to a 290X. Today I rearranged my GPUs so that my XFX 290 would be the secondary and this 290X would be the primary, thinking this would be a good way to get a little more oomph in games that don't handle CrossFire well. Isn't this the same XFX 290X giving you fits earlier? If so, I'd wager something is hosed with it.
|
|
#
?
Feb 1, 2016 08:35
|
|
|
FaustianQ posted:Isn't this the same XFX 290X giving you fits earlier? If so, I'd wager something is hosed with it. Turns out the problem with the thing was its BIOS, which as it was stock and unmodded, XFX's RMA department re-flashed it and sent it back. That being said they did it after saying "no, the card's fine". And they also replaced one of the fans because it wasn't spinning on their test bench, even though it spun fine in my system (in fact the only things that worked in my system on that card when the BIOS crapped out were the fans). That card is just possessed, I tell you what. e: Tomorrow I'm going to nuke all the drivers again and install clean ones. GPU-Z is reporting the XFX 290 as having the specs of the ASUS 290X and vice versa. e2: aaaaaand drivers crashed, twice. Tomorrow is going to be fun. Kazinsal fucked around with this message at 09:02 on Feb 1, 2016 |
|
#
?
Feb 1, 2016 08:56
|
|
|
So maybe you guys can help me. I've put a MSI Gaming 6G GTX 980 ti in my computer (i5-4690k, Asrock Z97e-ITX, 16GB Ballistix Sport) and performance isnt...good. I've run some heaven Benchmark on the same settings as these guys: https://www.reddit.com/r/nvidia/comments/39u5k4/post_results_of_your_980ti_score_in_heaven and i get about 50-60 FPS. I get that benchmarks aren't everything and that i wont get the same score as someone who does the whole exxxtreme OC and watercooling thing but 60 FPS? I can get 40 FPs with a GTX 970! I'm not new to building computers but im a bit stumped. Is there any kind of setting i could have missed? Or is my video card just broken?
|
|
#
?
Feb 1, 2016 09:08
|
|
Shai-Hulud posted:So maybe you guys can help me. I've put a MSI Gaming 6G GTX 980 ti in my computer (i5-4690k, Asrock Z97e-ITX, 16GB Ballistix Sport) and performance isnt...good. Check what clock speed the GPU is running at in Afterburner or something while running the benchmark. Also 40->60 FPS is about right for going from a 970 to a 980 Ti, it's a 50% increase which is just about how much more powerful the 980 Ti is than the 970.
|
|
|
#
?
Feb 1, 2016 09:20
|
|
|
Kazinsal posted:Turns out the problem with the thing was its BIOS, which as it was stock and unmodded, XFX's RMA department re-flashed it and sent it back. That being said they did it after saying "no, the card's fine". And they also replaced one of the fans because it wasn't spinning on their test bench, even though it spun fine in my system (in fact the only things that worked in my system on that card when the BIOS crapped out were the fans). Every single problem you've had so far has been this card, seriously. I'm doubting drivers and more sure that XFX support is garbage at helping you. Welp, looks like Sapphire and MSI for AMD cards for me, maybe some insane Powercolor SKU.
|
|
#
?
Feb 1, 2016 12:29
|
|
|
AVeryLargeRadish posted:Check what clock speed the GPU is running at in Afterburner or something while running the benchmark. Also 40->60 FPS is about right for going from a 970 to a 980 Ti, it's a 50% increase which is just about how much more powerful the 980 Ti is than the 970. I'll check clock speed when i'll get home tonight. Good idea. I mean, yeah it is a perforormance improvement, its just weird that everyone in that reddit thread is getting 20 fps more than me even with otherwise worse components. Maybe its overheating? Its pretty fast to go over 80C and i think i've read somewhere that 83C is the limit?
|
|
#
?
Feb 1, 2016 13:28
|
|
|
Shai-Hulud posted:I'll check clock speed when i'll get home tonight. Good idea. Also check that your windows power settings are not set to power saving
|
|
#
?
Feb 1, 2016 14:13
|
|
|
FaustianQ posted:Every single problem you've had so far has been this card, seriously. I'm doubting drivers and more sure that XFX support is garbage at helping you. Welp, looks like Sapphire and MSI for AMD cards for me, maybe some insane Powercolor SKU. I'm trying not to goon-in-the-well about this, but ugh. At this point I'm willing to just remove the XFX 290 and see if the ASUS 290X works fine on its own. Perhaps it's a lack of power supply oomph, but it'll crash while idle with the XFX in ULPS... I just want to be running fine and fairly well powerful until Polaris or Pascal. I shouldn't have tried crossfire again. loving hell.
|
|
#
?
Feb 1, 2016 16:41
|
|
|
Shai-Hulud posted:So maybe you guys can help me. I've put a MSI Gaming 6G GTX 980 ti in my computer (i5-4690k, Asrock Z97e-ITX, 16GB Ballistix Sport) and performance isnt...good. 40 to 60fps is a pretty big bump.
|
|
#
?
Feb 1, 2016 17:14
|
|
|
Oh also, after all the TR chat it looks like they had a patch before launch but after the first press build, and that patch is what techpowerup and pcper reviewed, and they have AMD performing better. Like 390 and 980 trading blows and the 970 strictly below the 390 like god intended better. What is up with that game. Feeling good about my 290 seeing those numbers and similar for the division though.
|
|
#
?
Feb 1, 2016 18:51
|
|
|
Shai-Hulud posted:So maybe you guys can help me. I've put a MSI Gaming 6G GTX 980 ti in my computer (i5-4690k, Asrock Z97e-ITX, 16GB Ballistix Sport) and performance isnt...good. 20 fps lower in Heaven is a lot. I would run DDU and go again. Don't have anything running but Heaven (even chrome will lower this score). Use afterburner as stated to see where you really stand on the numbers . With that set of info you can make better guesses about what's happening. I'll run heaven tonight, I have the same card. Oh and make sure you aren't frame limiting 
|
|
#
?
Feb 1, 2016 18:59
|
|
|
MikusR posted:Also check that your windows power settings are not set to power saving Ok yeah this was it. I did a fresh Windows 10 install and it put itself on energy saver mode. After disabling all that stuff i get 80 fps easy. Oh i also had to force my board to use PCIE 3.0 but i figured that one out earlier. Thanks for the help guys, im glad i didn't blow a shitton of money on a disappointing card!
|
|
#
?
Feb 1, 2016 21:04
|
|
|
I found setting windows to "Performance" but allowing CPU throttling is a great setup if you're gaming. Even "Balanced" is way too conservative and basically tries to downclock at every opportunity. "Performance" makes it prefer clocking up with any somewhat significant activity and I think it reduced little stutters and whatnot some in my case.
|
|
#
?
Feb 1, 2016 22:45
|
|
|
Shai-Hulud posted:
What was this about 
|
|
#
?
Feb 2, 2016 00:16
|
|
|
Fresh out the oven, AMD 16.1.1 "Hotfix" (not calling them betas now?) driver: http://support.amd.com/en-us/kb-articles/Pages/AMD-Radeon-Software-Crimson-Edition-16.1.1-Hotfix-Release-Notes.aspx Highlights Performance/Quality improvements and an AMD Crossfire™ profile is available for Rise of the Tomb Raider™ An AMD Crossfire™ profile is available for Fallout 4 e: At a quick glance it seems to have boosted my 7970 ~4fps in Tomb Raider, and turning off tessellation in game now bumps it up another ~5fps. Think I'm gonna be a big dweeb and wait a few months for my next card before playing it though so I can crank everything. teh_Broseph fucked around with this message at 01:10 on Feb 2, 2016 |
|
#
?
Feb 2, 2016 00:58
|
|
|
Dweeb nothing, I think that if you didn't get a new card in November/December, (barring any future hardware failure) you should hold off until the new poo poo comes out.
|
|
#
?
Feb 2, 2016 01:26
|
|
|
Anime Schoolgirl posted:What's most important is that they have good laptop SKUs--something AMD has been lacking for far too long.
|
|
#
?
Feb 2, 2016 01:42
|
|
|
Shai-Hulud posted:Oh i also had to force my board to use PCIE 3.0 but i figured that one out earlier. 
|
|
#
?
Feb 2, 2016 02:01
|
|
|
SemiAccurate thinks NVIDIA is significantly behind on Pascal. As in - the January Zauba shipments everyone freaked about were probably tools for the rev.01 engineering samples. Meanwhile, AMD showed a real, functional chip with real drivers and stuff in December. Again, I wouldn't be surprised if NVIDIA has to do a couple takes to get an HBM arch right. AMD has the advantage of a chip ready to shrink as soon as the process is ready. This probably means that anything shown in April will be paper bullshit or nonfunctional chips/mockups. However, please rest assured that upon launch Pascal will be the yoogest, most luxurious architecture money can buy. 
Paul MaudDib fucked around with this message at 03:18 on Feb 2, 2016 |
|
#
?
Feb 2, 2016 03:10
|
|
|
So quickly did Demerjian abandon his notions that NVidia had Pascal silicon ready for over a year!  If you're reading this, Charlie, we saw what you did, and we didn't buy in like chumps! =P
|
|
#
?
Feb 2, 2016 03:26
|
|
|
SwissArmyDruid posted:So quickly did Demerjian abandon his notions that NVidia had Pascal silicon ready for over a year! I ctrl-F'd through the source of the last 10 pages for semiaccurate, any link on that? I do remember one weird article from him but he seems like a weird dude so it might not have been that. SA gets a really eclectic and well-educated mix for an internet forum. There's about 80% less dumbassery here than on average. Y'all should pay attention to Duriana in particular, I'm reasonably versed in computer architecture (coming at it from compsci/GPGPU) and s/he's right on from what I know/follow. Please post more, Duriana 
Paul MaudDib fucked around with this message at 03:50 on Feb 2, 2016 |
|
#
?
Feb 2, 2016 03:43
|
|
|
Paul MaudDib posted:SemiAccurate thinks NVIDIA is significantly behind on Pascal. As in - the Zauba shipments everyone freaked about were probably tools for the rev.01 engineering samples. Meanwhile, AMD showed a real, functional chip with real drivers and stuff. SemiAccurate's conclusions are basically irrelevant since we all know what they're going to be, the important thing is that Charlie didn't have to resort to speculation. That, and that he's got a solid rationale. SwissArmyDruid posted:So quickly did Demerjian abandon his notions that NVidia had Pascal silicon ready for over a year! You mean the one where he was taking the piss the whole time and saying that NV had to have had Pascal silicon ready for over a year, before TSMC had figured out the process just waiting there and somehow not needing revisions because they wouldn't lie in an announcement, would they? The one where he ends with "[Jen-Hsun would never do something like that, right?" xthetenth fucked around with this message at 04:05 on Feb 2, 2016 |
|
#
?
Feb 2, 2016 03:57
|
|
|
Paul MaudDib posted:I ctrl-F'd through the source of the last 10 pages for semiaccurate, any link on that? I do remember one weird article from him but he seems like a weird dude so it might not have been that. http://semiaccurate.com/2016/01/11/nvidia-pascal-over-a-year-ahead-of-1416nm-competition/ And in case it goes bye-bye, 
|
|
#
?
Feb 2, 2016 04:01
|
|
|
xthetenth posted:You mean the one where he was taking the piss the whole time and saying that NV had to have had Pascal silicon ready for over a year, before TSMC had figured out the process just waiting there and somehow not needing revisions because they wouldn't lie in an announcement, would they? Paul MaudDib posted:CSI magic reveals the date codes on the chips on the Drive PX2 shown at CES to be over a year old. Feelin pretty good right now. Piss status: not taken. 
|
|
#
?
Feb 2, 2016 04:04
|
|
|
Oh drat. I did not realize you did that independently.
|
|
#
?
Feb 2, 2016 04:14
|
|
|
xthetenth posted:Oh drat. I did not realize you did that independently. Maybe it's all a meta-troll and I'm the AMD fanboi all along  spoiler: I did read the ExtremeTech article linked in my post but sorry Charlie that article did not make sense, and I can't tell the difference between super wrong and super sarcastic on this poo poo anymore double-spoiler: I am actually a super AMD fanboi since the K6-2 days but now own Intel/NVIDIA because I am not stupid Paul MaudDib fucked around with this message at 04:24 on Feb 2, 2016 |
|
#
?
Feb 2, 2016 04:18
|
|
|
Paul MaudDib posted:
I think that describes a lot of people around here.
|
|
#
?
Feb 2, 2016 04:42
|
|
|
El Scotch posted:I think that describes a lot of people around here. Yep, first full build was a K6-2 500. Stayed with AMD for CPUs until a couple of years ago. ATi/AMD video and Nvidia have traded between whatever was the best price/perf at the time though.
|
|
#
?
Feb 2, 2016 04:51
|
|
|
Price being the big variable as we ride out the 28nm doldrums. I'm running a "used" 390 I got for less than $200 a few months ago that was practically brand new because the owner absolutely had to have a 980ti.
|
|
#
?
Feb 2, 2016 05:15
|
|
|
Paul MaudDib posted:SemiAccurate thinks NVIDIA is significantly behind on Pascal. As in - the January Zauba shipments everyone freaked about were probably tools for the rev.01 engineering samples. Meanwhile, AMD showed a real, functional chip with real drivers and stuff in December. Hmmm...it's usually about 8 months from engineering samples to shipment, assuming nothing goes wrong, correct? It kind of sounds like Nvidia may want to consider a limited GDDR5X Maxwell refresh if anything goes wrong, an at best August release puts them in hot water compared to AMD and possibly puts consumer availability of their parts into 2017, that'd be a serious kick in the balls. Honestly, if I was Nvidia if I had to respin I'd keep doing it and just admit to it, rather than force out a potential Fermi.
|
|
#
?
Feb 2, 2016 05:47
|
|
|
|
| # ? May 30, 2024 23:59 |
|
|
Seamonster posted:Price being the big variable as we ride out the 28nm doldrums. I'm running a "used" 390 I got for less than $200 a few months ago that was practically brand new because the owner absolutely had to have a 980ti. Yeah, there's definitely performance overlap between AMD and NV, so even if AMD doesn't have quite as good parts right now, they can compete very well on price. I'm quite fond of my 290, it's a real solid competitor to the 970.
|
|
#
?
Feb 2, 2016 05:48
|
|