Guru3D Thermal Paste Roundup 2019
Back-to-school Microsoft Office 2016 $18 Win 10 Pro $12
TeamGroup T-Force Vulcan 1TB SSD Review
Corsair K57 RGB Wireless keyboard review
AMD Ryzen 7 3800X review
PowerColor 5700 RED DEVIL review
MSI Radeon RX 5700 XT Evoke review
ASUS Radeon RX 5700 XT ROG STRIX review
ViewSonic Elite XG240R Monitor Review
Promo: URCDKey Summer Sale Windows 10 Pro for $11
Nvidia Talks About Higher OC clocks on the Founder 2080 cards - also PCB Photo
NVIDIA just posted a blog post, and it is a rather unusual one. In the blog post, they pretty much state that they've always had an issue with the founders' edition cards, to reach higher boost clocks as to what the board partners have been able to achieve.
NVIDIA claims that with the new design a 90 MHz higher factory overclock is possible. That at least is the claim, and they also mention that each card is tested at the factory to ensure rock-solid OC stability which is backed by a 3-year standard warranty.
We've always had our comments about the blower style coolers from NVIDIA, they allowed a thermal threshold of 80 degrees, and all cards reached that in return throttling down.
The new two fan (13 blades) cooler design should be much better in that respect. And as long as the cards do not hit a too high-temperature target, of course, they will not throttle down, and in fact can even be clocked higher. NVIDIA mentioned that the new cooler will be much more silent as well, which is something I've complained about a lot with last-gen cooler. The new claim is that the cards will produce just a fifth of the acoustics compared to the last-gen design coolers, they mention it to be 8 DBa lower in noise, and that's a good value alright. We do wonder if NVIDIA isn't pushing too hard, as they are starting to really compete with their own AIB partners this way.
In other news, in the same blog-post, NVIDIA also posted a PCB shot and shows a photo of the new vapor chamber which covers the TU104 GPU (RTX 2080). GeForce RTX 2080 uses 225 Watts of power out of the box, and tops out around 280 Watts for enthusiasts chasing overclocking performance.
The power delivery system has been rebuilt for GeForce RTX Founders Edition graphics cards, starting with the all-new 13-phase iMON DrMOS power supply. Of particular note here is a new ability to switch off arbitrarily phases, for reduced power consumption at low workloads, which increases power efficiency.
NVIDIA Teases Heavily in Riddle Video - Deciphered - it is GeForce GTX 2080 - 08/14/2018 08:39 AM
NVIDIA has released a new video, a bit of a teaser called be for the game. If you look deeper and closer at the video it is FILLED with hints, and unofficially, it reveals the name of the to be releas...
NVIDIA Tweet Indicates New Volume Stock availability for GeForce graphics cards - 05/10/2018 01:19 PM
One more news-item from the NVIDIA front, the Nvidia team on twitter posted a tweet that suggests there is a large amount of graphics cards (stock) inbound. Which is good news, as prices should come ...
NVIDIA today announces NVIDIA Quadro GV100 GPU with NVIDIA RTX technology, - 03/27/2018 06:41 PM
NVIDIA today announced the NVIDIA Quadro GV100 GPU with NVIDIA RTX technology, delivering for the first time real-time ray tracing to millions of artists and designers....
NVIDIA to Host World's Top AI Experts at 2018 GTC - 03/05/2018 08:57 AM
NVIDIA will host thousands of the world's leading AI experts at its ninth annual GPU Technology Conference (GTC) on March 26-29 at the San Jose McEnery Convention Center. NVIDIA founder and CEO Jense...
NVIDIA To Launch GeForce GTX 1060 with 5GB Graphics Memory - 12/26/2017 06:56 PM
For a while now there has been some chatter on the web that Nvidia would be releasing or at least preparing a new GeForce GTX 1060, it seems the mystery has been solved. Nvidia is to launch a new GeF...
CalculuS
Senior Member
Posts: 2996
Joined: 2014-07-28
Senior Member
Posts: 2996
Joined: 2014-07-28
#5576383 Posted on: 08/22/2018 12:30 AM
The cooler design makes me think of EVGA ACX coolers.
The cooler design makes me think of EVGA ACX coolers.
RavenMaster
Senior Member
Posts: 1048
Joined: 2009-08-19
Senior Member
Posts: 1048
Joined: 2009-08-19
#5576391 Posted on: 08/22/2018 01:05 AM
You're a little selective when it comes to remembering your insults i see. Whatever, keyboard warrior. Good luck pulling that shit in real life
well, you would be considered pretty much brain dead if you bought one without seeing proper benches. Talk about stating the obvious.
You're a little selective when it comes to remembering your insults i see. Whatever, keyboard warrior. Good luck pulling that shit in real life
pimpineasy
Senior Member
Posts: 219
Joined: 2007-01-04
Senior Member
Posts: 219
Joined: 2007-01-04
#5576458 Posted on: 08/22/2018 06:41 AM
what is amd solutions slash vega to clearance they cant even beat 1080? ill buy it $100 rebranding gcn terascale is disgusting when product can only play lowest quality like gaming laptop 2018 that could heat a small house. rx was a flop rebrand for 60 fps min at 1080 rez. i got a 570 it very weak i call this minimum spec for gaming all it is rebadged r9 285 and it still a space heater. only plebs left without tensor cores and RT cores; amd will have physx performance on new titles if developer favor nvidia sdks over their own. pretty sure unreal is the industry which look good but all very similar. Looking forward for benches amd cant even enable tessellation yet, and nvida claims 6x better performance over 1080 for complex computation like RT. that probably means you can turn it on but the fps hit at high rez and frame times still might be trash.
what is amd solutions slash vega to clearance they cant even beat 1080? ill buy it $100 rebranding gcn terascale is disgusting when product can only play lowest quality like gaming laptop 2018 that could heat a small house. rx was a flop rebrand for 60 fps min at 1080 rez. i got a 570 it very weak i call this minimum spec for gaming all it is rebadged r9 285 and it still a space heater. only plebs left without tensor cores and RT cores; amd will have physx performance on new titles if developer favor nvidia sdks over their own. pretty sure unreal is the industry which look good but all very similar. Looking forward for benches amd cant even enable tessellation yet, and nvida claims 6x better performance over 1080 for complex computation like RT. that probably means you can turn it on but the fps hit at high rez and frame times still might be trash.
Robbo9999
Senior Member
Posts: 1289
Joined: 2012-10-07
Senior Member
Posts: 1289
Joined: 2012-10-07
#5576467 Posted on: 08/22/2018 07:32 AM
Nothing wrong with the guy surmising performance based on specs, it's just that...surmising. It's interesting to speculate, nothing wrong with it.
Did you just "surmise" the 2080ti performance from a pure CUDA core number count analysis? Just wow.
Nothing wrong with the guy surmising performance based on specs, it's just that...surmising. It's interesting to speculate, nothing wrong with it.
Fox2232
Senior Member
Posts: 9528
Joined: 2012-07-20
Senior Member
Posts: 9528
Joined: 2012-07-20
#5576482 Posted on: 08/22/2018 08:49 AM
Did you buy your terascale RX570 form China for $100? No wonder you write what you did. You should have bought 14nm Polaris based card not 28nm Tonga.
Well done trash post.
Your Athlon X4 CPU combined w/ GTX1050Ti speaks miles for your understanding.
what is amd solutions slash vega to clearance they cant even beat 1080? ill buy it $100 rebranding gcn terascale is disgusting when product can only play lowest quality like gaming laptop 2018 that could heat a small house. rx was a flop rebrand for 60 fps min at 1080 rez. i got a 570 it very weak i call this minimum spec for gaming all it is rebadged r9 285 and it still a space heater. only plebs left without tensor cores and RT cores; amd will have physx performance on new titles if developer favor nvidia sdks over their own. pretty sure unreal is the industry which look good but all very similar. Looking forward for benches amd cant even enable tessellation yet, and nvida claims 6x better performance over 1080 for complex computation like RT. that probably means you can turn it on but the fps hit at high rez and frame times still might be trash.
Did you buy your terascale RX570 form China for $100? No wonder you write what you did. You should have bought 14nm Polaris based card not 28nm Tonga.
Well done trash post.
Your Athlon X4 CPU combined w/ GTX1050Ti speaks miles for your understanding.
sdamaged99
Senior Member
Posts: 2025
Joined: 2006-12-12
Senior Member
Posts: 2025
Joined: 2006-12-12
#5576498 Posted on: 08/22/2018 10:02 AM
You can estimate teraflops performance from memory speed / bandwidth and clockspeed. CUDA cores are not really part of that equation. I'm not the only one who has done this, a few notable Tech YouTubers have also done the same and came to similar conclusions
This was from Forbes (who may be quoting someone else)
"Seriously, glance at the clock speeds for the 20 Series. Check out the unimpressive CUDA core increase over the 10 Series. Realize that the memory configuration is the same as the 10 Series (albeit with GDDR6 instead of GDDR5). Take a hard look at what the performance increase should be. Most in the tech media are putting it at maybe 10% to 15% over the 10 Series when it comes to the majority of games out there. But you'll pay 40% to 50% higher prices for this generation's replacements based on MSRP. And you know we won't be paying MSRP. . ."
All Nvidia talked about was Raytracing. Nothing on performance in games. Another red flag is a month of preorders, with no performance benchmarks allowed in the meantime. It's possible that the proprietary raytracing stuff will just die a death, as it will likely absolutely murder framerates similar to other gameworks technologies. Lets see
Did you just "surmise" the 2080ti performance from a pure CUDA core number count analysis? Just wow.
You can estimate teraflops performance from memory speed / bandwidth and clockspeed. CUDA cores are not really part of that equation. I'm not the only one who has done this, a few notable Tech YouTubers have also done the same and came to similar conclusions
This was from Forbes (who may be quoting someone else)
"Seriously, glance at the clock speeds for the 20 Series. Check out the unimpressive CUDA core increase over the 10 Series. Realize that the memory configuration is the same as the 10 Series (albeit with GDDR6 instead of GDDR5). Take a hard look at what the performance increase should be. Most in the tech media are putting it at maybe 10% to 15% over the 10 Series when it comes to the majority of games out there. But you'll pay 40% to 50% higher prices for this generation's replacements based on MSRP. And you know we won't be paying MSRP. . ."
All Nvidia talked about was Raytracing. Nothing on performance in games. Another red flag is a month of preorders, with no performance benchmarks allowed in the meantime. It's possible that the proprietary raytracing stuff will just die a death, as it will likely absolutely murder framerates similar to other gameworks technologies. Lets see
Evildead666
Senior Member
Posts: 1096
Joined: 2003-09-14
Senior Member
Posts: 1096
Joined: 2003-09-14
#5576502 Posted on: 08/22/2018 10:21 AM
So aside from color choices...what's the value of buying an AIB card over a Founder's Edition now that Nvidia is putting out cards with open air dual fan coolers? This feels like Nvidia is now looking to cannibalize their own AIBs and eventually cut them out completely. All they need to do is start offering water cooled versions of their cards at +$150 and the AIBs are trading solely based on their brand names.
This feels like preparation for Nvidia to take the whole thing in-house. I don't really have any strong feelings one way or the other on that, but it seems obvious moves are being made...
3DFX tried to cut out the Board partners.
Then Nvidia bought all their remains.
Nvidia knows all too well what happens when you cut out the AIB crowd.
They won't step on their toes too much, but I suspect people have been asking for High Build Quality cards straight from Nvidia, and they Obliged.
So aside from color choices...what's the value of buying an AIB card over a Founder's Edition now that Nvidia is putting out cards with open air dual fan coolers? This feels like Nvidia is now looking to cannibalize their own AIBs and eventually cut them out completely. All they need to do is start offering water cooled versions of their cards at +$150 and the AIBs are trading solely based on their brand names.
This feels like preparation for Nvidia to take the whole thing in-house. I don't really have any strong feelings one way or the other on that, but it seems obvious moves are being made...
3DFX tried to cut out the Board partners.
Then Nvidia bought all their remains.
Nvidia knows all too well what happens when you cut out the AIB crowd.
They won't step on their toes too much, but I suspect people have been asking for High Build Quality cards straight from Nvidia, and they Obliged.
metagamer
Senior Member
Posts: 1130
Joined: 2018-04-10
Senior Member
Posts: 1130
Joined: 2018-04-10
#5576599 Posted on: 08/22/2018 03:16 PM
You can estimate teraflops performance from memory speed / bandwidth and clockspeed. CUDA cores are not really part of that equation. I'm not the only one who has done this, a few notable Tech YouTubers have also done the same and came to similar conclusions
This was from Forbes (who may be quoting someone else)
"Seriously, glance at the clock speeds for the 20 Series. Check out the unimpressive CUDA core increase over the 10 Series. Realize that the memory configuration is the same as the 10 Series (albeit with GDDR6 instead of GDDR5). Take a hard look at what the performance increase should be. Most in the tech media are putting it at maybe 10% to 15% over the 10 Series when it comes to the majority of games out there. But you'll pay 40% to 50% higher prices for this generation's replacements based on MSRP. And you know we won't be paying MSRP. . ."
The 970 matched the 780ti and it had a lot less CUDA cores, it had a 256bit memory bus *cough* vs a 384bit memory bus therefore memory bandwidth was nowhere near. It ran around 200mhz faster than the 780ti though. And it matched the 780ti more often than not.
New architecture can also be a lot more efficient you see. So less CUDA cores and lower core clocks doesn't necessarily mean lesser performance. Let's not forget that the slower core clocks will be offset by higher memory clocks, to a certain point.
You can estimate teraflops performance from memory speed / bandwidth and clockspeed. CUDA cores are not really part of that equation. I'm not the only one who has done this, a few notable Tech YouTubers have also done the same and came to similar conclusions
This was from Forbes (who may be quoting someone else)
"Seriously, glance at the clock speeds for the 20 Series. Check out the unimpressive CUDA core increase over the 10 Series. Realize that the memory configuration is the same as the 10 Series (albeit with GDDR6 instead of GDDR5). Take a hard look at what the performance increase should be. Most in the tech media are putting it at maybe 10% to 15% over the 10 Series when it comes to the majority of games out there. But you'll pay 40% to 50% higher prices for this generation's replacements based on MSRP. And you know we won't be paying MSRP. . ."
The 970 matched the 780ti and it had a lot less CUDA cores, it had a 256bit memory bus *cough* vs a 384bit memory bus therefore memory bandwidth was nowhere near. It ran around 200mhz faster than the 780ti though. And it matched the 780ti more often than not.
New architecture can also be a lot more efficient you see. So less CUDA cores and lower core clocks doesn't necessarily mean lesser performance. Let's not forget that the slower core clocks will be offset by higher memory clocks, to a certain point.
fantaskarsef
Senior Member
Posts: 10588
Joined: 2014-07-21
Senior Member
Posts: 10588
Joined: 2014-07-21
#5576600 Posted on: 08/22/2018 03:19 PM
The 970 matched the 780ti and it had a lot less CUDA cores, it had a 256bit memory bus *cough* vs a 384bit memory bus therefore memory bandwidth was nowhere near. It ran around 200mhz faster than the 780ti though. And it matched the 780ti more often than not.
New architecture can also be a lot more efficient you see. So less CUDA cores and lower core clocks doesn't necessarily mean lesser performance. Let's not forget that the slower core clocks will be offset by higher memory clocks, to a certain point.
Just to add to that, iirc Maxwell had better compression algorythums than the generations before, that's how they were able to cope with less bandwidth for the memory.
The 970 matched the 780ti and it had a lot less CUDA cores, it had a 256bit memory bus *cough* vs a 384bit memory bus therefore memory bandwidth was nowhere near. It ran around 200mhz faster than the 780ti though. And it matched the 780ti more often than not.
New architecture can also be a lot more efficient you see. So less CUDA cores and lower core clocks doesn't necessarily mean lesser performance. Let's not forget that the slower core clocks will be offset by higher memory clocks, to a certain point.
Just to add to that, iirc Maxwell had better compression algorythums than the generations before, that's how they were able to cope with less bandwidth for the memory.
metagamer
Senior Member
Posts: 1130
Joined: 2018-04-10
Senior Member
Posts: 1130
Joined: 2018-04-10
#5576603 Posted on: 08/22/2018 03:26 PM
Yes, you're right. You never know, Turing might just have some wizardry up it's sleeve too. But just looking at numbers and determining performance solely like that is silly so I don't know why people do it.
Just to add to that, iirc Maxwell had better compression algorythums than the generations before, that's how they were able to cope with less bandwidth for the memory.
Yes, you're right. You never know, Turing might just have some wizardry up it's sleeve too. But just looking at numbers and determining performance solely like that is silly so I don't know why people do it.
Denial
Senior Member
Posts: 12235
Joined: 2004-05-16
Senior Member
Posts: 12235
Joined: 2004-05-16
#5576605 Posted on: 08/22/2018 03:31 PM
Turing has double the L1/L2 cache as Pascal which is going to alleviate hits to memory along with Variable Shading and Texture-Space Shading, both of which should make the overall process more efficient for memory bandwidth. And then yeah, whatever changes they made to delta compression.
Turing has double the L1/L2 cache as Pascal which is going to alleviate hits to memory along with Variable Shading and Texture-Space Shading, both of which should make the overall process more efficient for memory bandwidth. And then yeah, whatever changes they made to delta compression.
sdamaged99
Senior Member
Posts: 2025
Joined: 2006-12-12
Senior Member
Posts: 2025
Joined: 2006-12-12
#5576606 Posted on: 08/22/2018 03:33 PM
If it is indeed faster though, why not show some benchmarks? It seems like this has been a deliberate move, which doesn't bode well
If it is indeed faster though, why not show some benchmarks? It seems like this has been a deliberate move, which doesn't bode well
Denial
Senior Member
Posts: 12235
Joined: 2004-05-16
Senior Member
Posts: 12235
Joined: 2004-05-16
#5576610 Posted on: 08/22/2018 03:36 PM
Because I don't think it's that much faster. The only comparison they showed in regular workloads was with DLAA vs TAA in the infiltrator demo - where the framerate was doubled. You'll probably get some other titles that utilize DLAA, from what I understand it isn't that hard to implement, but aside from that I only expect a 25-30% performance uplift over 1080Ti with default clocks. Perhaps with overclocking - 2ghz, etc it will start to shine but idk. Like I said in the other thread, this entire series is going to be value-add features. The NGX stuff looks like an avenue for Nvidia to add a bunch of features to the card over its lifespan.
If it is indeed faster though, why not show some benchmarks? It seems like this has been a deliberate move, which doesn't bode well
Because I don't think it's that much faster. The only comparison they showed in regular workloads was with DLAA vs TAA in the infiltrator demo - where the framerate was doubled. You'll probably get some other titles that utilize DLAA, from what I understand it isn't that hard to implement, but aside from that I only expect a 25-30% performance uplift over 1080Ti with default clocks. Perhaps with overclocking - 2ghz, etc it will start to shine but idk. Like I said in the other thread, this entire series is going to be value-add features. The NGX stuff looks like an avenue for Nvidia to add a bunch of features to the card over its lifespan.
fantaskarsef
Senior Member
Posts: 10588
Joined: 2014-07-21
Senior Member
Posts: 10588
Joined: 2014-07-21
#5576612 Posted on: 08/22/2018 03:39 PM
I agree. I think we will only see what Turing really is worth for us if you look at it's performance compared to Pascal in DX11 (which won't do miracles I'm afraid). DLAA / DLSS is just a method of reducing workload in the GPU by using an approximative algorythum instead of brute calculations, hence that's how they got such a boost in that scenario. They compared it to TAA though, which because of the performance hit isn't the usual "gamer's choice". I too think we need to see overclocking performance too, since that's what "we" games will run, and that's where the real worth of an upgrade will be determined, not if you need RTX / DXR or not.
Because I don't think it's that much faster. The only comparison they showed in regular workloads was with DLAA vs TAA in the infiltrator demo - where the framerate was doubled. You'll probably get some other titles that utilize DLAA, from what I understand it isn't that hard to implement, but aside from that I only expect a 25-30% performance uplift over 1080Ti with default clocks. Perhaps with overclocking - 2ghz, etc it will start to shine but idk. Like I said in the other thread, this entire series is going to be value-add features. The NGX stuff looks like an avenue for Nvidia to add a bunch of features to the card over its lifespan.
I agree. I think we will only see what Turing really is worth for us if you look at it's performance compared to Pascal in DX11 (which won't do miracles I'm afraid). DLAA / DLSS is just a method of reducing workload in the GPU by using an approximative algorythum instead of brute calculations, hence that's how they got such a boost in that scenario. They compared it to TAA though, which because of the performance hit isn't the usual "gamer's choice". I too think we need to see overclocking performance too, since that's what "we" games will run, and that's where the real worth of an upgrade will be determined, not if you need RTX / DXR or not.
Click here to post a comment for this news story on the message forum.
Member
Posts: 23
Joined: 2017-11-24
There is absolutely no way that Nvidia would break from the AIB partners. Are we all already forgetting the Nvidia Partner Program? There is no way Nvidia could build cards fast enough on their own to keep up with initial demand for the consumer cards, as well as building all the Titans and Quadros for other markets as they currently do. When would they have time or resources for R&D???