Quick News



I think a LOT of people are going to be waiting for true gaming benchmarks in non RT games to assess the value of these.

It was definitely a little fishy that all this 6 x faster and so on was pushed toward games using RT. An idiot could've told you that RT games will benefit massively from Hardware designed to process it...
 
I think a LOT of people are going to be waiting for true gaming benchmarks in non RT games to assess the value of these.

It was definitely a little fishy that all this 6 x faster and so on was pushed toward games using RT. An idiot could've told you that RT games will benefit massively from Hardware designed to process it...

6x faster means 6x faster at RT than previous Pascal GPUs. We kinda knew that 'cause Pascal has no tensor cores specifically for the task.

However I did think that RT would destroy hardware and take a long time before we have GPUs powerful enough to really throw it around at big resolutions. If this is the 2080ti or two then just think how unfit for the task lower tiered cards will fare.

Yeah, not so good.
 
NVIDIA upgrades L1 and L2 caches for Turing

For Turing NVIDIA doubled the bandwidth for the L1 cache. Architectural design changes improved the latency and capacity of L1 cache, which is now 2.7 times as big as Pascal’s.

The L2 cache has been upgraded to 6MB, which is 3MB more than Pascal.

This slide was presented at Editors’ Day yesterday in Cologne. You are not supposed to see it till September 14, so don’t tell anyone.

Slide: https://videocardz.com/newz/nvidia-upgrades-l1-and-l2-caches-for-turing
 
with Nvidia supposedly making more of an effort to support Multiple GPUs with NVLink do you think we will see an improvement in SLI or do you think it's dead?
 
with Nvidia supposedly making more of an effort to support Multiple GPUs with NVLink do you think we will see an improvement in SLI or do you think it's dead?

I don't know about Turing yet (I don't think any one will with public info) but by now Nvidia should be reaching "scaleable" status. IE what AMD promised with Navi.

IDK if AMD will be using IF to stitch it all together or something else, but we should be reaching the point where they can make smaller dies and then tie them together. This makes them cheaper and easier to manufacture.

It's not SLi I am pretty convinced of that. Am sure we will find out once Nvidia want to sell it to us tbh.
 
with Nvidia supposedly making more of an effort to support Multiple GPUs with NVLink do you think we will see an improvement in SLI or do you think it's dead?

With Nvidia pushing RTX, it is likely that we will see more devs transition to DX12 or Vulkan, which passes the multi-GPU support almost entirely onto them. At least more so than DX11.

Nvidia and AMD can help push devs to support it with partnerships, but that won't really help much. It remains a niche audience within the PC gaming market, which typically means that the dev time is probably better placed somewhere else.
 
I've seen higher prices than that when you factor in VAT. This means Ryzen+ probably won't see a price drop since they're still competitively priced against whatever refresh-lake this new CPU range is.
 
Steam is currently letting people have For Honor for free for a limited time, to keep, I got my copy yesterday

I got mine to. Didn't download it though.

Played it on PlayStation. Game died quick but it's come back a little. Enough to wear you don't wait long for games at least

There was an article I read that after they announced it being free for the weekend the concurrent player count jumped from a few thousand to around 220k if I remember the number correctly. That was just steam alone. So the total was probably around 300k if you include Uplay
 
if those prices are real, Intel is dreaming in another world without AMD in it, holy balls what a ripoff.

Intel beating AMD in multi core performance i.e 8c/16 vs 8c/16t is basically already a guarantee and single core performance will be slightly better than the 8700K which is already faster than the 2700X by quite a bit so Intel can charge a premium for having the faster product but if it's THAT much more expensive then Intel are solely relying on brand loyalists.
 
Intel beating AMD in multi core performance i.e 8c/16 vs 8c/16t is basically already a guarantee and single core performance will be slightly better than the 8700K which is already faster than the 2700X by quite a bit so Intel can charge a premium for having the faster product but if it's THAT much more expensive then Intel are solely relying on brand loyalists.

The money saved into AMD platform can be spent on a better GPU and make the CPU irrelevant anyway and therefore end up with a faster PC anyway.
 
Back
Top