yeah you could even get a newer CPU I guess and just use the Intel iGPU encoder to not tax the AMD GPU but I’m used to Nvidia where the encoder hardware and the shader hardware are totally separate
this is… yet another respect where AMD is even further behind Nvidia than I thought post-2013
I would argue that Relive is better than Shadowplay because you don’t have to make an account to actually use the feature but if that bugs you enough, OBS has NVENC support
once you get into the RX series of cards, AMD’s encoder support is actually pretty much equal, it’s just the R9 2XX and HD cards that have less options for configuring the Relive software
oh yeah I’m like dunderheadedly cross-platform, it doesn’t even occur to me to use the official Windows kits they provide to access nvenc
MSI Afterburner for videocapture with MJPEG codec is your best bet, without any hardware encoding. It’s pretty light on CPU. Especially if you drop the quality a bit to like 90-93%
Also try Bandicam. It tends to capture more smoothly than afterburner. If your gpu has any hardware encoding, bandicam will use it.
I know you’re referring to Nvidia here but I feel the need to point out that AMD’s suite of Windows stuff is so hilariously comprehensive and on point that it’s embarrassing how bare Nvidia’s stuff is in contrast. their most recent drivers have OC, voltage and fan/temp controls, game recording, streaming and in-game overlay performance metrics standard. anything mentioned there that isn’t recording and streaming, you need an outside utility for Nvidia on Windows.
this isn’t even getting into how AMD is exposing a lot of this stuff and enabling quality open-source drivers on Linux
still waiting on that Sea Islands GPU support though, Linux kernel devs
sure but if I’m going to use ffmpeg anyway I’d rather have a better GPU
hey
Vega’s are pretty good
if you can find them
and they aren’t a thousand dollars
I just can’t bring myself to buy a card that’s 2/3 as efficient as the competition, at any price point. performance per watt is too important
what about Nvidia being evil
they send me free GPUs
being bribed, eh
like I’m in favour of open standards obviously but building a GPU and an interface to that GPU is massively massively complex and, demonstrably, many other corporations cannot do it, so to the extent that Nvidia de facto controls the market it’s not that surprising … things like GSync/Freesync are annoying obviously but it’s difficult to make the case that they should not have any proprietary technology, I don’t think much of their behaviour could be described as actively and unmitigatedly anticompetitive so much as the result of modern corporate welfare and increasing manufacturing complexity centralizing undue power in a single company much as it does in other sectors; that needs checks on it, but as far as Nvidia qua Nvidia, idk
and providing free stuff to the education/research sector is A-OK in my book obviously
also I keep wanting to call that AMD architecture “thousand island” which is a salad dressing
maybe if they put paul newman’s face on it in place of the bikini ninja
Unofficial reports are in that Nvidia is attempting to do to AMD graphics----what Intel did do to AMD CPUs and got in trouble for.
Both camps have always had some neat features which the other does not.
would that be “not accept proposed changes to the instruction set that they, in fact, maintain?” I’m not quite sure what would be analogous.
while I don’t begrudge proprietary standards, which forced into existence open standards, locking consumers into an ecosystem and upcharging on features to that extent comes off as gross
I would describe the (alleged) GPP as Actually Really Anti-competitive, seeing as it’s entirely an effort to (allegedly) control the consumer discrete graphics space through something as trivial as branding (allegedly) when they already have 80%+ market share
thousand island dressing is tasty and I won’t have Ruby defamed in such a manner
yeah those are all good points. Also worth noting that AMD had a straight up better consumer product than Intel from 2002-2005 when they really had to fight back, and even without that I’d be in favour of heavily funding competitors to balance the market, there just isn’t really a precedent for that, and it shouldn’t come with an expectation of consumers buying less efficient GPUs
really what should happen is, like, an anti-Nvidia consortium bloc, which is actually the effective reality w/r/t Linux drivers, that’s just not a big enough market to change anything, and the dynamics of open source driver development are not such that this can somehow produce a better product then Nvidia on their own
Because this is the extent to which tech power structures are reinforced
Reminds me of the Tizen consortium bloc that valiantly battled iOS and Android
It’s Arguable that AMD is only just now coming back financially, from the effects of what Intel did to them.
And if true bout Nvidia, the timing lines up. As AMD just reported a 40% earnings gain from last year.
Kyle at Hardocp doesn’t pull punches (famously hard on AMD through the FX CPU years and still hard on them for not competing with GPU at the top) and has already had a few things to say about GPP. Here is his latest:
Are folks breaking the First Rule of NVIDIA GPP: Don’t Talk About GPP? Allen Ngo of Notebook Check is reporting that he has spoken with “three independent and reliable sources” that are telling him that “Nvidia is strongly responsible for keeping Kaby Lake-G from proliferating.” While we have not spent a lot of time discussing Kaby Lake-G (the Intel mobile CPU with AMD’s Palo Alto GPU on substrate), we have touched on it a bit. Allen is flat out saying that NVIDIA is attempting to keep the Kaby Lake-G out of the market. One thing that we have not said publicly is that we believe that Kaby Lake-G was specifically targeted by another GPP-ish “program” before NVIDIA rolled out GPP, called “100% NVIDIA.” We have not discussed this, simply because we could not verify that with multiple sources, so I have sort of left 100% NVIDIA in the rumor bin. However, if what Allen is saying over at Notebook Check is true, then it would make me think that it is possible that a “100% NVIDIA” program actually exists. Guys talk, you hear things. Except for NVIDIA, they are still not saying a word about GPP, except when they try to disparage others talking about GPP.