NVIDIA GeForce RTX 5070 Ti Edges Out RTX 4080 in OpenCL Benchmark

[GGforever]

A recently surfaced Geekbench OpenCL listing has revealed the performance improvements that the GeForce RTX 5070 Ti is likely to bring to the table, and the numbers sure look promising - that is, coming from the disappointment of the GeForce RTX 5080, which manages roughly 260,000 points in the benchmark, portraying a paltry 8% improvement over its predecessor. The GeForce RTX 5070 Ti, however, managed an impressive 248,000 points, putting it a substantial 20% ahead of the GeForce RTX 4070 Ti. Hilariously enough, the RTX 5080 is merely 4% ahead, making the situation even worse for the somewhat contentious GPU. NVIDIA has claimed similar performance improvements in its marketing material, which does seem quiet plausible.

Of course, an OpenCL benchmark is hardly representative of real-world gaming performance. That being said, there is no denying that raw benchmarks will certainly help buyers temper expectations and make decisions. Previous leaks and speculations have hinted at a roughly 10% improvement over its predecessor in raster performance and up to 15% improvements in ray tracing performance, although the OpenCL listing does indicate the RTX 5070 ti might be capable of a larger generational jump, neck-and-neck with NVIDIA's claims. For those in need of a refresher, the RTX 5070 Ti boasts 8960 CUDA cores paired with 16 GB of GDDR7 memory on a 256-bit bus. Like its siblings, the RTX 5070 is also rumored to face "extremely limited" supply at launch. With its official launch less than a week away, we won't have much waiting to do to find out for ourselves.



View at TechPowerUp Main Site | Source

 

[Vayra86]

4%

This must be with some distance the laziest most shitty Nvidia gen ever. At least in Ada every GPU in the stack had a defined position in it and they all beat Ampere on perf/w. Today, you don't even have that, the better half of the stack is just a plain replacement for something you already had.

 

[GodisanAtheist]

Thanks to the dumpster fire situation with card pricing, we're not going to get 5080 performance for less, we'll end up with 5070Ti performance for 5080 prices.

Joy.

 

[scooze]

there was a result in blender on 5080 and it was fully confirmed. there is already a result 5070ti - 7521 points
+7% from 4070tis
4080super faster by 11%
closer to 4070tis than 4080

 

[Bomby569]

meh

 

[Chrispy_]

Thanks to the dumpster fire situation

SHHHHHHH!
Please don't use the word "fire" in the same sentence as 50-series. 600W down the 12VHPWR uh, I mean the 12V6X2 connector is perfectly fine and not in any way a problem. Everything is fine.

 

[Hyderz]

nvidia now realizes their mistake when they launched the gtx10 series, such a big increase in performance
then came 20 series and their performance over 10 series was rather minimal too.. 30 series had a slightly bigger boost...
i dont think we gonna see another repeat of 10 series perfomance again... kinda like intel's sandy bridge..

 

[Tpanon]

This must be with some distance the laziest most shitty Nvidia gen ever. At least in Ada every GPU in the stack had a defined position in it and they all beat Ampere on perf/w. Today, you don't even have that

I get that you are disappointed. But let's at least stick to the facts. According to TPUs review the 5080 is the most energy efficient card in the world. So the perf/w comment is perhaps a bit misplaced here.

NVIDIA GeForce RTX 5080 Founders Edition Review

The NVIDIA GeForce RTX 5080 is out now. Priced at $1000 it includes all the new GeForce RTX 50 features and offers good gaming performance. On the other hand, the gen-over-gen performance increase is smaller than expected, but NVIDIA didn't raise their pricing either.

www.techpowerup.com

 

[Rightness_1]

Who cares if you can't buy it?

 

[b1k3rdude]

I dont know who this info is supposed to benefit? as its an extreme edge niche case.

For everyone else the 4080 will still be faster than ther 5070ti across the board...

 

[Gatt]

... there is already a result 5070ti - 7521 points
+7% from 4070tis
4080super faster by 11%
closer to 4070tis than 4080

Oh boy, my poor 4070 Ti Super scores 7363 points. No reason to think about a 5070 Ti

 

[lexluthermiester]

"Edges" out? It should be beating it by a healthy margin

 

[rattlehead99]

nvidia now realizes their mistake when they launched the gtx10 series, such a big increase in performance
then came 20 series and their performance over 10 series was rather minimal too.. 30 series had a slightly bigger boost...
i dont think we gonna see another repeat of 10 series perfomance again... kinda like intel's sandy bridge..

The big increase in performance came entirely from them skipping 2 process nodes and 16nm being so great it carried them hard.
Pascal(gtx 1000 architecture) had literally 0 IPC improvement over Maxwell(gtx 900 architecture) and the changes were minimal, especially for raster.

The gtx 1000 series were the first rip-off generation, the 1080 is a 60 class card, sold as 80 class and at higher than ever prices(for the time).

 

[lexluthermiester]

Pascal(gtx 1000 architecture) had literally 0 IPC improvement over Maxwell(gtx 900 architecture) and the changes were minimal, especially for raster.

The gtx 1000 series were the first rip-off generation, the 1080 is a 60 class card, sold as 80 class and at higher than ever prices(for the time).

There's an old saying where I come from, and you're full of it..

 

[Rightness_1]

There's an old saying where I come from, and you're full of it..

Is it your job to sh't post attacks on people in this forum? Snide comments and personal attacks are all you even seem to post. I just responded to yet another of your personal attacks on me in another threat a few minutes ago.

I think the mod team should look at you.

 

[lexluthermiester]

Is it your job to sh't post attacks on people in this forum? Snide comments and personal attacks are all you even seem to post. I just responded to yet another of your personal attacks on me in another threat a few minutes ago.

I think the mod team should look at you.

Stop spewing senseless drivel and people won't call you out on it. Just throwing it out there.

Pascal(gtx 1000 architecture) had literally 0 IPC improvement over Maxwell(gtx 900 architecture) and the changes were minimal, especially for raster.

This was not only incorrect, but blatantly so. It shows your complete ignorance to factual info.

The gtx 1000 series were the first rip-off generation, the 1080 is a 60 class card, sold as 80 class and at higher than ever prices(for the time).

As does this. The GTX1000 series GPUs are very, very nearly universally praised for the excellent performance and value NVidia brought to the table. Only silly people, with little grip on reality, disagree.

So, you were saying?

I think the mod team should look at you.

Right back at you. Yes, yes.

 

[R0H1T]

4%

This must be with some distance the laziest most shitty Nvidia gen ever. At least in Ada every GPU in the stack had a defined position in it and they all beat Ampere on perf/w. Today, you don't even have that, the better half of the stack is just a plain replacement for something you already had.

From Toms ~

Japanese RTX 50 retailer apologizes for lottery-fueled mayhem — unruly crowd caused injuries, damaged nearby kindergarten

Store and neighboring businesses were overwhelmed by crowds.

www.tomshardware.com

Now imagine if they could do this in Japan, then for the rest of the world

 

[rattlehead99]

Stop spewing senseless drivel and people won't call you out on it. Just throwing it out there.


This was not only incorrect, but blatantly so. It shows your complete ignorance to factual info.

As does this. The GTX1000 series GPUs are very, very nearly universally praised for the excellent performance and value NVidia brought to the table. Only silly people, with little grip on reality, disagree.

So, you were saying?


Right back at you. Yes, yes.

Ah yes they skipped 2 process nodes(22nm and 20nm) and then they sold you a 60 class GPU for 600-700$ (FE)MSRP, they called it a gtx 1080, when it was a 1060 and priced it higher than ever before. And the performance is great because again they skipped 2 process nodes. That's like going from the gtx 1000 series on 16nm to the RTX 6000 series, that isn't even out yet, on 3nm and praising how much faster the rtx 6000 series are compared to the 1000 series. Of course they are.

Pascal literally has no IPC improvement over maxwell, the differences are minimal. There are even videos showing the gtx 1050TI vs gtx 950 which have the same specs, just different clock speeds, because 16nm was a lot better than 28nm, when clocked at the same core and memory clock they perform identical.

The gtx 1000 series were carried by 16nm and they literally did what they have been doing with the rtx 4000 and 5000 series.