NVIDIA’s RTX Speed Claims “Fall Short,” Ray Tracing Merely “Hype”

You're assuming perfect scaling. 4K is 4x's the resolution of 1080p. Not 2x's. So you would need 4x RTX2080Ti to reach those kind of resolutions ASSUMING if it scaled well which SLI setups rarely (IF EVER) do.
Um, I said it was ~4x. If you bothered to read, you'd notice I included projections for 2070, 2080, 2080Ti, and 2x 2080Ti. The 2x 2080Ti was doubling of a 1440p projection, which is approx. correct, not the 2070 1080p projection. Go re-read it.

Edit: the thread where I originally made this prediction, has all the math laid out.
 
Um, I said it was ~4x. If you bothered to read, you'd notice I included projections for 2070, 2080, 2080Ti, and 2x 2080Ti. The 2x 2080Ti was doubling of a 1440p projection, which is approx. correct, not the 2070 1080p projection. Go re-read it.

Edit: the thread where I originally made this prediction, has all the math laid out.

*chuckles* First off you are still assuming 100% scaling which is WRONG. The best scaling I remeber is about 50% benefit with 1 additional card. Even NVIDIA has claimed SLI is dead. And they are moving away from it for a reason. It's antiquated. And 1440p is HALF the pixel resolution of 4K. So 1.5x's on 2x's the resolution under the best conditions isn't going to cut it.

Second early benchmarks show well below 30fps with RTX ON on 4K using the 99% percentile on frame times. Doubling your speed would still make you consistently hit below 60 which is the modern bar for games @ 4K
 
The clock on that leak was 2025MHz IIRC. It's OC a substantial amount. Back that out, assume linear, and it was inline with core scaling projections.

What is of concern here is the lower base clock limit. Those additional RTX circuits do produce heat and energy consumption even IF they aren't in use. So the question is how does clock speed scale compared to a Ti side by side. Does it throttle more?

That questions cannot be answered till we have reviews in. I'm personally expected about 30% increase in raw fps speed with similar settings compared to previous gen. So MAYBE 50% premium cost for 30% speed.

Overall it's way too early to tell without reviews. But RTX is in it's infancy and as is, may cost too much resource wise to use practically. It's more of a side-step in terms of performance. You have to trade off one (resolution) for visual enhancement on a few games.

I'm hoping for the best, but I'm not expecting miracles. Ignore the marketing hype and wait for Kyle's review.
 
What is of concern here is the lower base clock limit. Those additional RTX circuits do produce heat and energy consumption even IF they aren't in use. So the question is how does clock speed scale compared to a Ti side by side. Does it throttle more?

That questions cannot be answered till we have reviews in. I'm personally expected about 30% increase in raw fps speed with similar settings compared to previous gen. So MAYBE 50% premium cost for 30% speed.

Overall it's way too early to tell without reviews. But RTX is in it's infancy and as is, may cost too much resource wise to use practically. It's more of a side-step in terms of performance. You have to trade off one (resolution) for visual enhancement on a few games.

I'm hoping for the best, but I'm not expecting miracles. Ignore the marketing hype and wait for Kyle's review.
I don't know and it sounds like we won't know for a few more weeks.

I pointed out the OC because it's affecting the overall assessments of performance w/o RT. It seemed to me that most comparisons across the web left out the fact it was heavily OC and were comparing the score against stock cards when claiming % increase.
 
*chuckles* First off you are still assuming 100% scaling which is WRONG. The best scaling I remeber is about 50% benefit with 1 additional card. Even NVIDIA has claimed SLI is dead. And they are moving away from it for a reason. It's antiquated. And 1440p is HALF the pixel resolution of 4K. So 1.5x's on 2x's the resolution under the best conditions isn't going to cut it.

Second early benchmarks show well below 30fps with RTX ON on 4K using the 99% percentile on frame times. Doubling your speed would still make you consistently hit below 60 which is the modern bar for games @ 4K
Yes, in the original post where I did all the math, I explained all that.

No, I don't expect 100% scaling and didn't when I made that estimate. I do expect higher scaling than SLI as SLI always had inadequate bandwidth and latency beyond sharing a few frame buffers, etc. NVLINK is a different beast. It won't scale 100%, but I'm hopeful it's going to scale much better than SLI.

I think the problem is, my OP in this thread referenced another post where this was all laid out in detail and you simply haven't bothered to go read it. You're knocking down straw man arguments I didn't make.
 
Yes, in the original post where I did all the math, I explained all that.

No, I don't expect 100% scaling and didn't when I made that estimate. I do expect higher scaling than SLI as SLI always had inadequate bandwidth and latency beyond sharing a few frame buffers, etc. NVLINK is a different beast. It won't scale 100%, but I'm hopeful it's going to scale much better than SLI.

I think the problem is, my OP in this thread referenced another post where this was all laid out in detail and you simply haven't bothered to go read it. You're knocking down straw man arguments I didn't make.


SmartSelect_20180831-133030_Chrome.jpg
 
Hmmmm,
$600 for 2070 (new) or $350 shipped for used 1080 on ebay?

I would go with the GTX 1080 now and wait till next year when GPU prices are more reasonable

That's exactly what I did. I said, "Screw it. I'll get the 1080 and see where things are in a year or so."
 
Back
Top