RTX3080 vs 6800XT 50 game benchathon, Fine Wine is here?

[Grim5]

Hardware Unboxed has restested 50 games between the two and compared the result to how these two cards performed one year ago.

However, the relative results were the same - the performance difference between the 3080 and 6800XT is the same today as it was 12 months ago, there is far no evidence of fine wine for RDNA2 which in previous generations of GPUs allowed AMD to close the gap/extend the lead with Nvidia

 

[LtMatt]

HUB still physically disable Smart Access Memory, so the majority of their RDNA2 results are wrong, some by up to 20%.

 

[NZXT30]

12 months is not enough for the fine wine effect.

 

[Darujhistan]

I see HUB are still shilling for Nv.

 

[Wrinkly]

They really should do a RT comparison as a few here were claiming, just after launch, AMD would catch up over time.

I think this wine is corked, send it back lads

 

[willhub]

HUB still physically disable Smart Access Memory, so the majority of their RDNA2 results are wrong, some by up to 20%.

What about RBAR for NV though?

 

[tommybhoy]

What about RBAR for NV though?

Thats why it's disabled, its mostly a checkbox 'we can do it too' feature on Nv.

Disabling SAM on AMD, Rebar on Nv, skews the results, HUBs viewers are going to hammer them about it again.

 

[Ayahuasca]

What about RBAR for NV though?

Makes next to no difference in 99% of titles that support it.

 

[Nexus18]

Fine wine............

They really should do a RT comparison as a few here were claiming, just after launch, AMD would catch up over time.

I think this wine is corked, send it back lads

But in all seriousness, it is a bit of a flawed test for a couple of reasons:

- No SAM and no rebar, both see a boost in performance despite what some amd fans want you to believe. SAM does see more of a gain, largely in amd sponsored titles though
- no ray tracing tests

Funny no mention of vram issues there...... must be "nvidia shills!!!!"

Makes next to no difference in 99% of titles that support it.

Depends on the setup and game but on the whole, it does improve performance overall:

Several other videos showing perf. improvements too.

It made a huge difference on the 1% and 0.1% lows in cp2077 on my end.

 

[RavenXXX2]

Sam was enabled in forza.

 

[Ayahuasca]

Several other videos showing perf. improvements too.

It made a huge difference on the 1% and 0.1% lows in cp2077 on my end.

That video doesn't really sell it. Same performance in a lot of titles with it enabled and worse performance in some. Yeah, I overexaggerated with the 99% but it makes nowhere near as much difference as SAM does for AMD cards.

 

[Nexus18]

That video doesn't really sell it. Same performance in a lot of titles with it enabled and worse performance in some. Yeah, I overexaggerated with the 99% but it makes nowhere near as much difference as SAM does for AMD cards.

Yeah that's why I said it's worthwhile checking out other videos on youtube Some setups see much bigger gains, some see none. Not to mention, games and drivers from nvidia have probably improved over time since that video. Would be a good video that tbh, see if there has been any improvement.

Saying that I don't expect rebar to ever be as good as SAM since AMD are going to be able to get far more from their optimisations/drivers given amd cpu, amd based chipset/mb, amd gpu and consoles optimisation in this regard will probably carry over too.