AMD launches Instinct MI300X and MI300A CDNA3 accelerators

https://videocardz.com/press-release/amd-launches-instinct-mi300x-and-mi300a-cdna3-accelerators

Certainly a lot of claims by AMD but once you read the finer details, you do see some interesting quirks. The jist of it stems from two key areas, one of which is that AMD did not use TensorRT LLM with H100 which means it's not performing to it's full potential and the other is that using their benchmarks, it largely sounds like it will be 20% faster than H100 until H100 hits a vram limitation at which point MI300X will be much faster. Keep in mind that H100s released back in March and H200s will be releasing Q2 next year.

It is interesting because despite MI300X having 30% higher theoretical compute performance, we aren't really seeing the results which shows that AMD has a long ways away to optimize their software stack in the data center. You can read more about it here:

https://www.semianalysis.com/p/amd-mi300-performance-faster-than

With all that being said, I think AMD is still very much ahead of Intel in the Ai race which in itself is an achievement

AMD announces Ryzen 8045HS, 8040HS and 8040U “Hawk Point” series powered by Zen4, RDNA3 and XDNA

https://videocardz.com/newz/amd-announces-ryzen-8045hs-8040hs-and-8040u-hawk-point-series-powered-by-zen4-rdna3-and-xdna

AMD next-gen “Strix Point” series with XDNA2 NPU to offer 3x performance boost for generative AI

https://videocardz.com/newz/amd-next-gen-strix-point-series-with-xdna2-npu-to-offer-3x-performance-boost-for-generative-ai

I wonder if having hardware whether it be a CPU or GPU will be required as a minimum when it comes to installing Windows 12

Intel 144 Core Sierra Forest Xeon CPU Leaks Out: Powered By Sierra Glen E-Core Architecture, 172 MB Cache

https://wccftech.com/intel-144-core-sierra-forest-xeon-cpu-leak-e-core-architecture-172-mb-cache/

ARM-Based Chips From Qualcomm & Apple To Give AMD & Intel Tough Time In AI PC Segment In 2024

https://wccftech.com/arm-chips-qualcomm-apple-give-amd-intel-tough-time-in-ai-pc-segment-2024/

While I think Intels Core Truth marketing is dumb and over the top, AMD certainly isn't helping their case

Avatar: Frontiers of Pandora benchmarks

https://www-computerbase-de.translate.goog/2023-12/avatar-frontiers-of-pandora-benchmark-test/3/?_x_tr_sl=auto&_x_tr_tl=en&_x_tr_hl=en#abschnitt_benchmarks_in_full_hd_wqhd_und_ultra_hd

Keep in mind that this is one of those games that has Ray Tracing enabled regardless of setting so you can't turn off Ray Tracing. It is only software level RT instead of hardware RT as far as I can tell but it is still RT none the less hence why we see surprising lead for Nvidia gpus even though it is an AMD sponsored title. But the advantage of software is you can still play the game even on GTX/RDNA 1 cards.

FSR 3's later iteration has fixed VRR issues found in the previous two titles and this is one of the better FSR Super Resolution games although DLSS Super Resolution still takes the cake as being superior. The only issue with FSR 3 is how they handle the UI as the generated frame basically doesn't render the UI which means the UI is rendered at half rate which produces fairly noticeable artifacts around the UI. Not that DLSS 3s rendered UI produces that much better results either so both companies need to work on their UI rendering solutions.

A decent step up for AMD Instinct, a lot of it due to the smaller node and higher power budget, of course, but still competitive nevertheless.

Blackwell has to at least match its 192 GB RAM, right? It's so easy to choke a previous generation/competitor with a large enough inference model in these slides.

Probably as H200 is 143GB so B100 will likely be more than that. Of course with Nvidia nothing is free

I'm honestly expecting the worst from 12. They already showed that they felt we all wanted a massive UI change and even more settings hidden within more cryptic menus (which none of us really asked for). 

Them having ads won't be far off either. They'll likely use the AI to basically tap into what you type, what you do for daily activities, what you watch/listen to, and then slowly seep in ads to you over time, or worse, use the Ai to corroborate with ad makers into delivering their ads to you.

Either way the inclusion of AI comes off as a net negative, rather than a boon for the consumer. it feels strongly like it is designed primarily to benefit MS and anyone else they work with.

I know, because I don't need an AI to help me with anything I do as general activity. I don't need an AI to open my mail let alone read it, don't need it to update apps for me, don't need it to tell me what the weather forecast is like, don't need it to tell me what new tv shows or movies are out, and I ultra specifically do not need it to advertise to me what they are selling, be it their Xbox app or OS based sales. 

THe only aspect it could aid me with is media design, and I know full well that there isn't an AI atm that assists with that. The only chores AI can do atm is read or generate content it has to steal from existing sources to mash them all together, which is no help for someone like me at all. So all in all, the inclusion of AI into their OS feels like a dead weight I never asked for in an OS (also it'll take up processes that I didn't ask to be taken up on the CPU/memory either). 

CEO really thinks we all wanna lose the start button in favour of shitty voice commands, which factually take longer to announce than the user simply clicking one damn button?.

Can the tech industry just expunge the boomers already, they are really getting out of touch with what the user actually wants. I do not find voice commands "revolutionary" nor beneficial to me at all, especially seeing as how AI still cannot fully recognise different language accents (like Scottish/Liverpool/Manchester/Wales).

You're right though, this blunder is going to make 11 look way better than the shit they have planned for 12. 12 just looks like the worst user experience, and it's clear that those working on it are only thinking of themselves/one tiny sector of a market that none of us really care about.

That timeline is actually sad, not only because of how long their new games take to develop, but because of all the other good IPs that they've left behind to focus only on GTA and RDR... and judging by how quickly they ditched RDR2 and RDR Online despite how well it sold, it wouldn't surprise me if they only make GTA games from now on.

By the way, that list misses Lemmings. It's weird to think that that game and GTA are the creation of the same people.

Call me tinfoil, but after looking at that new timeline, I can't help but think that R* now just wants to string alone 1 game every decade now. Like when you think about how they treated the last 3 games of theirs (GTA V, RDR2 and GTA DE), they only treated one of theirs for longevity purposes (online play), but even then, all 3 in terms of SP were not really expanded upon.

Basically it feels like R* is trying to experiment with the live service template everyone else has been using, only they add some shiny content up front (them marketing the single player for all their games), and then just focus on what they really want for nearly a decade.

This is why I ended up losing interest in R*, because what they want isn't what I want. Yeah the visuals of GTA VI look great, but I haven't the hardware to run it, nor do I want to be treated like shit via waiting extra long for a PC port (that's magically going to be more demanding than the console ports, I still hate how they did this with GTA V and RDR2), and having to deal with a client that doesn't need to exist. 

Over time I started to realise that they ended up ditching RDR2 because T2 themselves had set expectations for it, as in they wanted it to sell on the level of GTA V, online and all, and because it didn't match 1-1 with GTA V, they just decided "eh good enough" and left it to rot (which is a live service way of thinking).