Nvidia's 3000 Series RTX GPU [3090s with different memory capacity]

Yup, this makes no sense if there was stock, but in the current reality I'm sure they'll all sell out instantly. 3090 looks like better value imo, compared to the 3080 Ti.
 
It's kinda funny seeing reviewers complaining about the MSRP while simultaineously saying that it won't be available even at that MSRP either.

Yes, it's a money grab by Nvidia. No, it would still be a money grab if the MSRP would be lower, just not by Nvidia. What's the difference? At least Nvidia can use these money to make better GPUs in the future. So I don't understand the complains.
 
Nvidia's Reflex Technology Beats AMD's Radeon Boost Counterpart: Report | Tom's Hardware (tomshardware.com)
March 20, 2021
Last year, in a bid to increase system responsiveness in competitive video games, both Nvidia and AMD released new technologies to reduce input lag for improved system performance: Nvidia's Reflex Low Latency and AMD's Radeon Boost. Both Nvidia and AMD say their respective technologies' performance gains are significant, but how true is that really?

Fortunately, YouTuber Battle(non)sense recently published a full rundown of both technologies and compared them side by side to see how they actually perform.
...
Thanks to Battle(non)sense, we've learned that optimizing the pipeline between the CPU and GPU is a much better way of lowering latency than simply pumping up the frame rate.

Plus, we've also learned that using a frame rate cap can significantly reduce input lag, enough to where it's on par with Nvidia's Reflex technology. That makes this solution absolutely perfect for the vast majority of titles that do not support Nvidia Reflex.
 

Why is Reflex so much better than Ultra Low Latency mode? In that mode there's basically no frame buffering so it must mean that the CPU is much faster than the GPU which is probably not the case with most gamers.

Reflex seems to mostly benefit lower end GPUs that max out at relatively low framerates. At higher framerates > 144Hz latency should be much less of a concern overall even if you're GPU bound.
 
Is Reflex still locked behing GeForce Experience which requires account registration for some god awful reason?
 
Why is Reflex so much better than Ultra Low Latency mode? In that mode there's basically no frame buffering so it must mean that the CPU is much faster than the GPU which is probably not the case with most gamers.

Reflex seems to mostly benefit lower end GPUs that max out at relatively low framerates. At higher framerates > 144Hz latency should be much less of a concern overall even if you're GPU bound.


Ultra low latency mode just doesn't work very well. The best way to reduce latency is to reduce your gpu load below 98% with a frame-rate limiter (preferably in-game) or use Nvidia Reflex which will allow you to push your gpu to 100%.
 
Any drawback from Reflex ? It's seems, on paper, it should be always enabled and the default way to make stuff work ?
 
Last edited:
I'd expect the usual - lower average fps and potentially higher CPU load.

You may get slightly lower fps than if you were to run reflex off with an uncapped framerate, but your input lag will be much lower. You're much better off taking the small hit to your frame rate and getting a more responsive game. I haven't seen any testing regarding cpu usage with relfex, but I can't see why it would use any significant amount of cpu time compared to reflex off.
 
Last edited:
I haven't seen any testing regarding cpu usage with relfex, but I can't see why it would use any significant amount of cpu time compared to reflex off.
Lower CPU queue depth may lead to higher CPU usage in some cases. But generally you're correct, these are minor things and I wouldn't even call them downsides.
 
That may be the case but I haven't seen a good explanation as to why it doesn't work.

I haven't seen any detailed explanation. ultra low latency is supposed to set the render queue to zero by forcing it at the driver level. Reflex aims to do the same thing, but integrated at the engine level. So I'm assuming by forcing the queue to zero outside of the scope of the rendering engine, there's some inefficiency in achieving the desired result.
 
Reflex takes the concept of Ultra Low Latency Mode a strep further, instead of being a driver level solution that tampers with the render queue, it becomes an engine level solution which tracks what the GPU is doing and synchronizes this with the CPU and the render engine, achieving even more response and latency reduction, it likely works across input devices as well.

The SDK is exclusive to NVIDIA GPUs, the feature also includes a boost mode that ensures that the GPU is operating at max boost even when massively underutilized, as in cases of massive CPU limitations.
 
It's kinda funny seeing reviewers complaining about the MSRP while simultaineously saying that it won't be available even at that MSRP either.

Yes, it's a money grab by Nvidia. No, it would still be a money grab if the MSRP would be lower, just not by Nvidia. What's the difference? At least Nvidia can use these money to make better GPUs in the future. So I don't understand the complains.

From my observations here, the bigger local retailers, who didn't have server problems serving the huge requests, are all out of stock. Smaller online retailers, some of them had to "restart" the sale, tend to have some still available in stock, though mostly those "overclocked" version which're sold at higher than MSRP prices (like, US$400 higher).
Also, some of the pre-built packages (bundled with CPU, MB, DRAM, SSD, etc.) are also in stock. They also tend to be priced slightly higher than the retailing prices of the components combined. But it's probably convenient for people who wants to upgrade most components of their current PC at the same time.
 
Back
Top