Morgan MLGman

Update 25 Oct. 2023
FSR3 Frame Generation was released early October. To try it out yourself for free you can download the Forspoken demo and enable it. As a game integration, this also works on nvidia GPUs.
AFMF has been released for testing in a preview driver, now supporting RDNA2 as well as RDNA3 GPUs.
 
Content below is the original news post and will not be further updated.
 
Summary
AMD's answer to NVIDIA's DLSS 3 frame generation has been demonstrated. Reports are it looks the equal to DLSS 3, but it was a hands-off demo so general people haven't had a chance to experience how it feels yet. Coming September 2023. Hardware recommendations for FSR 3 frame generation: Supported RX 5700, RTX 20 series, Recommended RX 6000 series and above, 30 series and above. 
 
An early demo of AMD Fluid Motion Frames (AFMF) was also shown. This is a "generic" frame generator that works at driver level reportedly working for DX11 and DX12 titles, but this comes at a cost of requiring at least RDNA3 GPU. This is expected Q1 2024.
 
Quotes
 
My thoughts
FSR 3 as a vendor agnostic implementation of frame generation could open up opportunities for more people. It is positive that those from the Digital Foundry team who looked at AMD's demo were happy with the visual output and comparable to nvidia's DLSS 3. We'll have to wait and see how the "feel" of it goes since that is still a questioned point for nvidia's implementation.
 
The further demo of AFMF is also interesting. Much like RSR gave a generic implementation of FSR in the driver, AFMF is FSR 3 in driver, but limited to RDNA3 GPUs. Does it rely on some tech only present in that generation or is it an early implementation thing, with wider support to come? Note AFMF in driver implementation will not have access to game engine vector information so it may be more comparable to a TV like interpolation with lower latency.
 
There are some practical considerations to using either of these, but that also applies to DLSS 3 anyway so isn't a new problem.
 
Edit: trying to understand some things. FSR 3 seems to be the combination of FSR + AFMF + Anti-Lag+, in a parallel that DLSS 3 is DLSS SR + FG + Reflex. What combination of those components will be used if FSR 3 will be used on non-AMD hardware? Could you for example swap Anti-Lag+ with Reflex for example? Anti-Lag+ is an improved implementation requiring Radeon 7000 series GPUs. Anti-Lag remains in use for RX 6000 series and older GPUs. So FSR 3 will be something that's going to take potentially more testing than DLSS 3 due to the wider permutations possible.
 
Sources
https://www.eurogamer.net/digitalfoundry-2023-amd-reveals-long-awaited-fsr-3-tech-and-frame-gen-for-every-dx11dx12-game
https://community.amd.com/t5/gaming/first-look-at-amd-fidelityfx-super-resolution-3/ba-p/626581
https://gpuopen.com/fsr3-announce/
 
 

AM5 can only hit 6000C30 for the most part especially without EXPO. On X3D its even harder and i see 5600C30/28 being better in stability for those CPU. Run it manually at 6200/6000 as @Morgan MLGman suggests, but id just return it for a 6000C30 kit with EXPO profile inside, itll be cheaper than your RGB corsair kit most likely.

Looking at it again, you're right, but I was partially right. It looks like the two x16 slots split the CPU lanes to x8+x8 if both are in use.
 
From the link I posted before:

 
It looks like if you put a 4 M.2 card in the main (normally GPU) slot with 2nd empty, that can run 3 M.2 drives. If you use the 1st slot as x8 (potentially one x4 M.2) then the 2nd slot can break its 8 lanes into x4+x4 supporting two more M.2 drives. It looks like the first slot can only break the lower 8 lanes as one group.

I guess so, although comparing yourself to an even worse offender rarely makes you look much better 😛 Nvidia's current prices are ridiculous, so RDNA 2 and RDNA 3 cards look great in terms of value. The 7900 XTX was supposed to be decent value but is stupidly expensive in this case, because it should retail for around $999. The 7900 XT's value was always questionable, so IMO it's either the top-end one or the 6950 XT which is priced very well considering its rasterization performance is on par with a 3090 Ti.
Normally I'd agree, however in this case it may not be that bad for the 4090. We're in an age of AI, ML and LLMs (Large language models) and owing to that, the 4090 may retain a lot of its value for longer than usual. This is obviously pure speculation on my part but I think it makes sense.

Honestly the 6950 XT is priced very well there - the 7900 XT's price is "meh" and the 7900 XTX is stupidly expensive.
If the 6950 XT's performance is enough for your needs, I wouldn't look any further as the recently released RTX 4070 is quite a bit slower for a similar price.

This the same guy that said RX6000 cards dies were cracking due to drivers? 
 
IMO this is a non-story. These cards have been out for over two years and preventing sag has well known prevention options. 
 
Fake drama is boring. 

The 7900 XTX hands down, not even a contest in this case.

The 7900 XTX hands down, not even a contest in this case.

Honestly the IPC uplift from the Zen+ architecture was like 2% at best, so considering the 1800X has 100MHz higher boost clock it's probably about the same. 1800X all the way for the same price, although for gaming even lower-end but newer Ryzen chips will be better than an 1800X or the 2600.

I saw some hoodies in Creator Warehouse last time I was over there. I will have to follow up and find out for you.

Obvious question, but are you sure that the ESXi server also has the same time zone set?

Here's a knowledge base article on how to change it if necessary: https://kb.vmware.com/s/article/1436

Thanks for the concern everyone. We are still in recovery mode over here and working with YouTube to get everything restored. Will hopefully have a video (or at least an update on WAN Show) to share with you all ASAP, but we want to make sure we get the details right since smaller channels may rely on our experience to help harden their own security.

6950 XT is certainly still a relevant GPU for modern gaming, and should be relevant for 1440p gaming for several more years. 
 
The 4070 ti performs 4% better than the 6950 XT on average, according to TechPowerUp's relative rankings. (The gap will grow with ray tracing enabled but you said you don't care about that.) Since in your local market the 4070 ti is ~50% more expensive, I don't see any case to be made for that... paying 50% more for a performance uplift that's barely outside the margin of error. 

6650XT is 13% faster than 3060, and 3060's RT performance is pitiful, FSR is almost as good as DLSS, then again, DLSS in 1080p is not recommended due to how blurry it is. Literally no reason to go with 3060 if you only game.
So, 6650 XT.

VRMs are designed to withstand temperatures up to about 120C for a very long period of time - temps when the GPU is idle or is around 60C are nothing for the VRMs. You've got nothing to worry about. They only get hot when there's a lot of power coming through them - in those instances the GPU is hot as well, so the fans spin up.

VRMs are designed to withstand temperatures up to about 120C for a very long period of time - temps when the GPU is idle or is around 60C are nothing for the VRMs. You've got nothing to worry about. They only get hot when there's a lot of power coming through them - in those instances the GPU is hot as well, so the fans spin up.

VRMs are designed to withstand temperatures up to about 120C for a very long period of time - temps when the GPU is idle or is around 60C are nothing for the VRMs. You've got nothing to worry about. They only get hot when there's a lot of power coming through them - in those instances the GPU is hot as well, so the fans spin up.

VRMs are designed to withstand temperatures up to about 120C for a very long period of time - temps when the GPU is idle or is around 60C are nothing for the VRMs. You've got nothing to worry about. They only get hot when there's a lot of power coming through them - in those instances the GPU is hot as well, so the fans spin up.

VRMs are designed to withstand temperatures up to about 120C for a very long period of time - temps when the GPU is idle or is around 60C are nothing for the VRMs. You've got nothing to worry about. They only get hot when there's a lot of power coming through them - in those instances the GPU is hot as well, so the fans spin up.

AMD has the 390X2 (2x R9 390) and Radeon ProDuo which are two Fury X chips on one GPU. Those GPUs are not "dead", Nvidia abandoned the idea I guess, though I'd expect AMD to release a dual Vega card this year.

the 'windows xp era' ones had some flair, but all the benefit that might have offered has been outdone by modern cases.
 
modern alienware cases are just a dell optiplex with a carnival store fat suit around them.

I meant in terms of build quality, internal design etc. Alienware cases are mostly just cheap, generic metal boxes found in stuff like Dell's Optiplex PCs, but wrapped around with tons of unnecessary plastic. At least the ones that they make now. Example below:


Notice how it's just a square made from cheap metal with giant plastic covers all around.

Actually the NH-D15 may be better than an AIO for good RAM temperatures - with an AIO you have a lot of space, but not as much air moving because there are no fans on the CPU block.

What motherboard and RAM kit do you have? Which DIMM slots are occupied in your setup currently?

It looks like your VBIOS is not up-to-date, as Nvidia added support for Resizable Bar later to be on par with AMD.

Your version is 94.04.25.40.41 from GPU-Z, and that is a version released at launch, and I think the latest one is 94.04.3A.40.63 that was released in March 2021. But you'd need to check specifically for your model to be sure, I believe you have a MSI Ventus RTX 3070 3X OC, but you need to verify that.

According to MSI, VBIOS can be updated in two ways:

I'd be very curious what the real reasons are behind this. If you take their statement at face value, does that mean the 4080 12GB just disappears? They must have made them to build up stock for launch. Some must already physically exist. So the question then is, are they going to re-launch it (later) as a 4070-something?