Chinese outlet Teclab leaks RTX 3090 benchmarks

[FaxedForward]

Teclab has leaked gaming benchmarks for the GeForce RTX 3090 ahead of the embargo, just like they did for the RTX 3080. Their earlier RTX 3080 benchmarks were accurate, so these figures should be accurate as well.

 

Wccftech overview:

Quote

 

NVIDIA GeForce RTX 3090 vs RTX 3080 (compilation by Videocardz)
Score / 4K AVG FPS	RTX 3090	RTX 3080	3090/3080
3DMark Time Spy Extreme	9948	9000	+10.5%
3DMark Port Royal	12827	11981	+7.1%
Metro Exodus RTX/DLSS OFF	54.4	48.8	+10.2%
Metro Exodus RTX/DLSS ON	74.5	67.6	+10.2%
Rainbow Six Siege	275	260	+5.8%
Horizon Zero Dawn	84	76	+10.5%
Forza Horizon	156	149	+4.7%
Far Cry	107	99	+8.1%
Assassins Creed Oddysey	71	65	+9.2%
Shadow of the Tomb Raider RTX/DLSS Off	91	83	+9.6%
Shadow of the Tomb Raider RTX/DLSS On	111	102	+8.8%
Borderlands 3	67.6	61.3	+10.3%
Death Stranding DLSS/RTX ON	175	164	+6.7%
Death Stranding DLSS/RTX OFF	116	104	+11.5%
Control DLSS/RTX ON	71	65	+9.2%
Control DLSS/RTX OFF	62	57	+8.8%

 

 

 

I see the 3090 as a Titan RTX successor and not a gaming card. The price increase over the 3080 is unjustified for gaming performance; you should only be shopping for a 3090 if you need the VRAM for professional workloads (or if money is no object).

 

Sources

Teclab video: https://www.bilibili.com/video/BV1dZ4y1N79a

Videocardz analysis of Teclab video: https://videocardz.com/newz/nvidia-geforce-rtx-3090-gaming-performance-review-leaks-out

Hardwaretimes analysis of Teclab video: https://www.hardwaretimes.com/nvidia-rtx-3090-gaming-benchmarks-leaked-just-10-faster-than-the-rtx-3080/

Wccftech analysis of Teclab video: https://wccftech.com/nvidia-geforce-rtx-3090-teclab-review-leaked/

[Moonzy]

10% faster for 100% more price

 

the VRAM might come in handy, though

[Applefreak]

Those numbers make little sense to me. Gamers Nexus did a live stream and overclocked some RTX 3080 cards on air, seeing almost the same results in Port Royal than the supposed 3090, makes me think they either had the wrong card or did something really wrong in their testing process. There is no way the performance is that bad.

But then they would have had to use a modified 3080 driver, likely not being able to address more cores on the chip so that makes sense I guess.

[FaxedForward]

6 minutes ago, Applefreak said:

But then they would have had to use a modified 3080 driver, likely not being able to address more cores on the chip so that makes sense I guess.

I'm thinking the hacked driver is definitely a contributing factor.

 

That said, I would also bet the last bit Wccftech mentions is relevant; the RTX 3080 has not even been out a full week, the software support for this new generation is probably just not there yet.

Quote

The first answer that comes to my mind is that the amount of core increase that we saw in the 3000 series is just too big for software stacks to handle. While the drivers would (probably) have been updated to handle the massive throughput, game code and engines have to scale up to take advantage of the available processing power as well. This is sort of like games being optimized primarily to take advantage of just 1 core and not scaling perfectly.

 

What we are seeing with the RTX 3080 and 3090 also appears to be a similar problem where the hardware is being bottlenecked by software. AMD's GPUs are usually fondly referred to as finewine but if my gut instinct is correct, the RTX 3000 series is going to turn out to be the biggest load of finewine silicon the gaming world has ever seen. With the cards only delivering half of the performance promised by NVIDIA I am fairly certain we are going to get massive incremental performance improvements pushed via software.

 

[pzspah]

Nvidia 3090 graphs could be quasi-fake to throw AMD off the scent. Unlikely but not impossible.

[TetraSky]

Quote

Death Stranding DLSS/RTX ON	175	164	+6.7%
Death Stranding DLSS/RTX OFF	116	104	+11.5%

 

Turning RTX off gives less fps than with RTX ON ???

Am I understanding this correctly? How does that work? I thought RTX was more taxing on the GPU, thus giving less fps overall? Did that change, for some reason? Is this real life?

[PopReference]

Just now, TetraSky said:

 

Turning RTX off gives less fps than with RTX ON ???

Am I understanding this correctly? How does that work? I thought RTX was more taxing on the GPU, thus giving less fps overall? Did that change, for some reason?

Death stranding only has DLSS not ray tracing so it's saying RTX features all on

[FaxedForward]

1 minute ago, TetraSky said:

 

Turning RTX off gives less fps than with RTX ON ???

Am I understanding this correctly? How does that work? I thought RTX was more taxing on the GPU, thus giving less fps overall? Did that change, for some reason?

Looks like they turned DLSS and RTX on at the same time. DLSS probably provided a net uplift.

[.Apex.]

36 minutes ago, Applefreak said:

Those numbers make little sense to me. Gamers Nexus did a live stream and overclocked some RTX 3080 cards on air, seeing almost the same results in Port Royal than the supposed 3090, makes me think they either had the wrong card or did something really wrong in their testing process. There is no way the performance is that bad.

But then they would have had to use a modified 3080 driver, likely not being able to address more cores on the chip so that makes sense I guess.

the 3090 is 20% faster than the 3080 theoretically, so it isn't outlandish to only see a 10% performance difference in the real world as it should never exceed the theoretical performance uplift

 

as an example the RTX 2080Ti is 48% faster than the 2080 spec wise, but it only saw a 20-30% performance uplift in games.

[.Apex.]

18 minutes ago, FaxedForward said:

I'm thinking the hacked driver is definitely a contributing factor.

NVIDIA released drivers for the 3080 and the 3090, I don't think there's any modification involved

[Fatih19]

Imagine if AMD released a GPU that is meant to compete with the 3080 but accidentally have a 3090-like performance because the difference between the two is not that big. That'd be funny.

[eeeee1]

17 minutes ago, Fatih19 said:

Imagine if AMD released a GPU that is meant to compete with the 3080 but accidentally have a 3090-like performance because the difference between the two is not that big. That'd be funny.

imagine how fast you could go on the amd bike... that would be very impressive 

[StDragon]

1 hour ago, Moonzy said:

10% faster for 100% more price

 

the VRAM might come in handy, though

Ahh yeessss! The law of diminishing returns. The Universe is at a balance once again.

[Moonzy]

1 hour ago, eeeee1 said:

imagine how fast you could go on the amd bike... that would be very impressive 

stability issues though, heard a lot of people crashed while using it

[eeeee1]

7 minutes ago, Moonzy said:

stability issues though, heard a lot of people crashed while using it

nvidia: haha amd, i have the high ground now!

amd: haha nvidia, bike goes brrrrrrrrrrrrrrrrrrrrrrrrrrrrr

nvidia: no amd, dont try it!

[eeeee1]

1 hour ago, Fatih19 said:

Imagine if AMD released a GPU that is meant to compete with the 3080 but accidentally have a 3090-like performance because the difference between the two is not that big. That'd be funny.

"There are no accidents."

-Master Oogway

[Suika]

Those are much more disappointing numbers than I anticipated, I was expecting a similar uplift in the 3090 to the 3080 as we saw with the 2080 Ti to the 2080, but that might have been unreasonable to expect given the spec differences between all of those cards.

 

I'll still be waiting for third party reviews to confirm these numbers before I jump the gun and grab a 3080 (and at this rate, wait for Big Navi's announcement), but if these numbers are even close to realistic, I'll be disappointed and will likely wait for that rumored 20GB 3080 SKU.

[CTR640]

I think it's the driver. They don't have the latest and newest driver yet for the 3090. nVidia is not stupid for that 10%, no one with a healthy common sense would pay up $1500+ for a 10% performance.

[Đỗ Đức Huy]

Meh, I should be getting mine in the mail any day now. The local distributor sure take their sweet time NOT seeding cards to end retailer. Lucky I have a laxed client who are fully endorsing me to test out the 3090 before giving it to him, All as thanks for helping him flipped the 2080ti days before the 3000 series launched.

[Suika]

27 minutes ago, CTR640 said:

I think it's the driver. They don't have the latest and newest driver yet for the 3090. nVidia is not stupid for that 10%, no one with a healthy common sense would pay up $1500+ for a 10% performance.

You're forgetting all the Titan cards that existed before Turing lol.

[CTR640]

23 minutes ago, Suika said:

You're forgetting all the Titan cards that existed before Turing lol.

Weren't the Titans supposed to be semi-pro?

[FaxedForward]

1 hour ago, CTR640 said:

Weren't the Titans supposed to be semi-pro?

Nvidia said in the reveal video that the 3090 is supposed to the the Titan’s replacement.

[suits]

This is definitely gonna need to be an official review before purchase product. 

[suits]

4 hours ago, CTR640 said:

I think it's the driver. They don't have the latest and newest driver yet for the 3090. nVidia is not stupid for that 10%, no one with a healthy common sense would pay up $1500+ for a 10% performance.

I'm pretty sure Nvidia said the performance was better than that too. I know Nvidia overhypes but not that much lol

[Valentyn]

I want stock!

Also let's see a about other reviewers. Did we get given a date? Maybe the 23rd?