Friday, January 3rd 2020

NVIDIA's Next-Generation Ampere GPUs to be 50% Faster than Turing at Half the Power

Jan 3rd, 2020 03:00 Discuss (227 Comments)

As we approach the release of NVIDIA's Ampere GPUs, which are rumored to launch in the second half of this year, more rumors and information about the upcoming graphics cards are appearing. Today, according to the latest report made by Taipei Times, NVIDIA's next-generation of graphics cards based on "Ampere" architecture is rumored to have as much as 50% performance uplift compared to the previous generations of Turing GPUs, while using having half the power consumption.

Built using Samsung's 7 nm manufacturing node, Ampere is poised to be the new king among all future GPUs. The rumored 50% performance increase is not impossible, due to features and improvements that the new 7 nm manufacturing node brings. If utilizing the density alone, NVIDIA can extract at least 50% extra performance that is due to the use of a smaller node. However, performance should increase even further because Ampere will bring new architecture as well. Combining a new manufacturing node and new microarchitecture, Ampere will reduce power consumption in half, making for a very efficient GPU solution. We still don't know if the performance will increase mostly for ray tracing applications, or will NVIDIA put the focus on general graphics performance.

Source: Taipei Times

Add your own comment

227 Comments on NVIDIA's Next-Generation Ampere GPUs to be 50% Faster than Turing at Half the Power

#176

efikkan

Super XPI was basing the results on UserBenchmark results. As they compare 1,000's of gaming PCs.

I don't care if it was every single PC in the whole world. Benchmarks like this are 100% useless. You need controlled test conditions, otherwise the test result is completely worthless.
I like Steve at Gamer's Nexus' comment about these pages; they are just search engine spam.

#177

lexluthermiester

efikkanYou need controlled test conditions, otherwise the test result is completely worthless.

That depends on the test and the conditions thereof.

efikkanI like Steve at Gamer's Nexus' comment about these pages; they are just search engine spam.

Steve gets a lot of things right, but like every other human being, he doesn't know everything.

#178

cucker tarlson

lexluthermiesterThat depends on the test and the conditions thereof.

that's exactly what he's saying

#179

EarthDog

cucker tarlson41%

cucker tarlson75%

I just think he doesn't get the math. He's right if you change one word. Where he says faster it should read slower as the graphs are based on the 980ti and 1080ti. The way you are doing it is based on how much faster the faster card is over the slower card. :)

or....its based on userbench??? Wth?? Oye.

#180

Super XP

cucker tarlson41%

75%

first you claim rt cores do nothing for rt and nvidia just gimped 1080ti to make rtx line look better
then you lower performance numbers by over 50%

I know it takes a lot to believe amd is still competitive but what you are writing is quite a stretch even for those standards.

I do admit the GTX 1080ti is one of the bests GeForce graphics cards Nvidia released in a very long time. Not even the 2080ti can compete with how well the 1080ti was priced for the performance and the actual gain you got versus the previous 980ti. In other words, no I am not impressed with the RTX line of GPU's, and I believe many people agree with me.

Thanks for the TPU review nevertheless,..

efikkanI don't care if it was every single PC in the whole world. Benchmarks like this are 100% useless. You need controlled test conditions, otherwise the test result is completely worthless.
I like Steve at Gamer's Nexus' comment about these pages; they are just search engine spam.

I can agree with that, as controlled benchmarks would show real performance improvement differences. As with non controlled can be somewhat manipulated to show altered results.

#181

medi01

A note about 980Ti.
Fury X beat it at 4k at release (1440p later on/depending on tests) was very close to it otherwise in 1080p (who cares, but oh well).

That was "countered" by a valid argument, that that is stock 980Ti and actual AIBs are OCable good 20% beyond that.

However, once 1080 popped out, people have forgotten about that nice quirk of 980Ti and now we in all seriousness discuss how 1080 was 33%-ish faster.

#182

Prima.Vera

50% faster on relative or actual performance??

#183

InVasMani

50% cheaper at 50% less power so seeing as it's Nvidia it'll cost what 200% more?

#184

Fluffmeister

Whatever their next gen cards bring, they will no doubt be the fastest most efficient cards available.

Sorry about that.

#185

InVasMani

FluffmeisterWhatever their next gen cards bring, they will no doubt be the fastest most efficient cards available.

Sorry about that.

Be realistic...:rolleyes:

#186

lexluthermiester

InVasMani50% cheaper at 50% less power so seeing as it's Nvidia it'll cost what 200% more?

Your logic, along with your math, is flawed.

#187

lewis007

CalmmoThis is talking about ampere, will nvidia even make gaming GPUs from that or are they again going to hold off for a year then release a cut down variant under a different name? AMD have nothing to show still, officially, so why would they unless the new node allows for a significant reduction in costs. If this is an indication of anything.. this is probably about their new titan 3000$ card..

Agree, 50% perf uplift 150% price uplift.

#188

BiggieShady

Prima.Vera50% faster on relative or actual performance??

Not sure if joking but if somehow serious, the answer is both, because that's how ratios work ... 50% faster is ratio of 1.5 meaning fasterSpeed = 1.5 * slowerSpeed means fasterSpeed is 50% faster than slowerSpeed. Relative or actual performance numbers, ratio stays the same, it's all relative in the end.

#189

efikkan

I wouldn't get too fixated on precise performance estimates. Until Nvidia have the final batch of qualification samples, no one really knows, including Nvidia themselves.

#190

Anymal

100% uplift at the same W! 3070 with 170w will be 100% faster than 2060. Also price, 650eur vs. 325eur.

#191

Fluffmeister

50% better performance per watt would be disappointing over Turing, exciting times ahead either way.

#192

TheoneandonlyMrK

Anymal100% uplift at the same W! 3070 with 170w will be 100% faster than 2060. Also price, 650eur vs. 325eur.

So 100% performance uplift at 325 Euro , I think both a little optimistic personally.

#193

medi01

I think it's also worth noting that 1080Ti price went way beyond 980Ti's price and even 1080 cost more than 980Ti.

If perf bumps are true, it will be reflected in pricing accordingly.

#194

Vayra86

medi01I think it's also worth noting that 1080Ti price went way beyond 980Ti's price and even 1080 cost more than 980Ti.

If perf bumps are true, it will be reflected in pricing accordingly.

Yes, but then if you look beyond one generational increase to the next, the price per 'Frame / FPS' is still going down every time. A bigger perf jump is only paid in full when you early adopt it. At the same time however, when performance jumps per generation stall like they did the past few years, you see the trend of price per frame stall as well. After all, cards still gotta get sold but the market is saturated with a certain performance level already. There is not much to compete over, so there is less competition.

#195

medi01

Vayra86Yes, but then if you look beyond one generational increase to the next, the price per 'Frame / FPS' is still going down every time

Yes, but nowhere as close as people comparing Ti vs Ti implied was my point.
1080Ti is not a correct comparison to much cheaper 980Ti, it's even arguable if 1080 is.

#196

EarthDog

InVasManiBe realistic...:rolleyes:

It isn't? Have you seen the current landscape? Nvidia is more efficient and faster in most cases. And that is at 12nm, not 7nm...

#197

Vayra86

medi01Yes, but nowhere as close as people comparing Ti vs Ti implied was my point.
1080Ti is not a correct comparison to much cheaper 980Ti, it's even arguable if 1080 is.

780TI MSRP: $699
980TI MSRP: $649
1080TI MSRP: $699

780TI > 980TI

980TI > 1080TI

Say what again now? 1080ti's price point was far too low given its performance. And let's face it, that price point is STILL relevant today...

2070S MSRP: $499

See how bigger perf jumps translate to bigger price gaps over time? Hell, the 2070S (=1080TI) even cannibalizes the 2080 and the 2080S for most people because of its price drop. Its a performance perf/dollar king now and that is also what the 5700XT is priced towards today (of course going under the price with slightly lower perf as usual).

To drive it home

1070 MSRP: $379

Complaining about the Pascal price hike was never realistic, and here is your proof. We got a LOT more for our money there than we ever did during Kepler or Maxwell. Especially in terms of VRAM. Everything's got a lush 8 GB (or more) to work with. The ONLY reason the 980ti is still relevant is because its the only Maxwell card with 6GB. Try that with the budget 970...

#198

medi01

Vayra861080TI MSRP: $699

Nice try.
Now let's check reality, shall we:

1080 (non TI) FE: $699
1080 (non TI) "later on edition": $599

Did I mention it was "non-TI"? Thanks.
Now, TI version came later, after 1080 milked the market for about 1 year.

#HowToMilk101

Vayra862070S MSRP: $499

S right? The one released after AMD spanked 2070?
That's cool.

2070 non S, however, released for $599 for FE edition.
Are you telling me it was slower than 1080Ti? Oh, what as strange "improvement" of perf/$, chuckle.

#199

EarthDog

medi01Nice try.
Now let's check reality, shall we:

1080 (non TI) FE: $699
1080 (non TI) "later on edition": $599

Did I mention it was "non-TI"? Thanks.
Now, TI version came later, after 1080 milked the market for about 1 year.

#HowToMilk101

S right? The one released after AMD spanked 2070?
That's cool.

2070 non S, however, released for $599 for FE edition.
Are you telling me it was slower than 1080Ti? Oh, what as strange "improvement" of perf/$, chuckle.

1080ti's msrp on release day was $699... look it up...so was the 1080 a few months prior.

Last I checked, they are a for profit business. Perhaps if amd had anything worthwhile at the time, prices wouldnt have inflated so much...

#200

Vayra86

medi01Nice try.
Now let's check reality, shall we:

1080 (non TI) FE: $699
1080 (non TI) "later on edition": $599

Did I mention it was "non-TI"? Thanks.
Now, TI version came later, after 1080 milked the market for about 1 year.

#HowToMilk101

S right? The one released after AMD spanked 2070?
That's cool.

2070 non S, however, released for $599 for FE edition.
Are you telling me it was slower than 1080Ti? Oh, what as strange "improvement" of perf/$, chuckle.

As always your strange pair of glasses made you handily gloss over a key word, competition. You are exactly right about the 2070S. That is how movement happens: when the competitor can challenge similar performance. So, if the contrary happens and performance stalls, while being unchallenged, MSRP is unlikely to drop. That was the initial point. No performance movement = no or very slow price movement.

And even without competition the performance cost was reduced by 100 bucks MSRP, so, your point?

Add your own comment

NVIDIA's Next-Generation Ampere GPUs to be 50% Faster than Turing at Half the Power

227 Comments on NVIDIA's Next-Generation Ampere GPUs to be 50% Faster than Turing at Half the Power

Latest GPU Drivers

New Forum Posts

Popular Reviews

Controversial News Posts

NVIDIA's Next-Generation Ampere GPUs to be 50% Faster than Turing at Half the Power

Related News

227 Comments on NVIDIA's Next-Generation Ampere GPUs to be 50% Faster than Turing at Half the Power

Latest GPU Drivers

New Forum Posts

Popular Reviews

Controversial News Posts