Quote from: Timothy Tripp on November 04, 2024, 15:59:58I would LOVE to see nVidia put out a great ARM chip to be used for gaming,

GeForce is not good enough? I mean, they are literally using ARM architecture for their GeForce GPUs...

Quote from: Hotz on November 05, 2024, 20:05:11Man... I'm tired of this capitalistic bullshittery...

A bit of history for those of you to young to have lived through it...

The "WINTEL" cartel has for decades orchestrated a 'planned obsolescence' of computer systems. The latest examples are the hardware 'requirements' for WIN11.

While there are SOME in the industry that would claim that the x86 'backwards compatibility' is a LIMITING factor in x86's development+growth, Intel's nearly constant platform changes makes a LARGE percentage of PC hardware incompatible generation to generation.

I applaud Nvidia entering the PC space with an APU however I'd guess that Nvidia would be up to their same old tricks - IE: creation of obvious software features that are ONLY available on next generation products.

Keep in mind that while the ONLY real difference between a VERY EXPENSIVE Nvidia 'professional GPU' card and a consumer card is the cards PCIe ID, NOBODY has ever brought suit to Nvidia for the outlandish practice.
 

Addendum to my post above:

Nvidia probably also dislikes, that on a CPU-GPU platform (APU) they can't fob you off with limited VRAM. Because the System RAM is your VRAM. So they can't sell you more VRAM for premium prices. Imagine this nightmare of Nvidia: that even low-end gamers could manually upgrade their "VRAM" by simply buying another 32gb RAM bar for 100$, and have no VRAM bottleneck anymore.

Because of that there is also a chance, that these future Nvidia APUs will only come with soldered RAM, so that you have no chance to upgrade yourself.

Interestingly enough it looks like AMD starts to think likewise: for example, I noticed that all devices with Strix Point APUs with 890m are using soldered RAM instead of DIMM. Coincidence? I doubt it. And why is still no 9000G or 10000G Desktop APU series planned? It feels like they're hesitating to give consumers too much flexibility. Almost as if the product is too good and low-cost for the common people.
Instead they focus on Halo - which will not be a budget product, but highly priced, and most likely will also come with soldered RAM only...

Man... I'm tired of this capitalistic bullshittery...

QuoteCPU-GPU platform which targets high-end PC market

And so it becomes uninteresting again. The high-end-PC market is already served well. The low-midrange-PC market is, where APU could really shine. Imagine socketed APUs (like 8000G) from Nvidia... we certainly would have better APUs now. And imagine you could simply upgrade your PC by simply only buying a new socket processor....


But... all too good to be true. They wouldn't allow us to upgrade a PC simply like that. Nvidia would play the same tricks as it does now:
"oh, you want an APU with better Nvidia iGPU? Time to upgrade your mainboards because socket changes."
"oh, you want to use our new DLSS5 technology? Buy new APU, because it's blocked on your old APU."

I can see Nvidia faring better than Qualcomm.

It's funny how much people disdain Intel in here, and they casually forget both Qualcomm & Nvidia are worse companies toward consumers. Really few people upgrade yearly, so idk why people complain so much about "muh lack of support" for their current platform. Say that if there are issues that don't get resolved with BIOS updates or firmware updates and I can bet you neither Nvidia or Qualcomm will allow you to upgrade anything at all.

Do you want to pay whatever price Nvidia asks out of their CPUs in the case they do capture the market like they did with GPUs?
Have you guys forgotten that Qualcomm has released the same cpu under three different names (SD 690, 695, 6 Gen 1 , etc) during 3 years?
And nvm their lack of interest for long term support for their previous flagship chips...

Quote from: Timothy Tripp on November 04, 2024, 15:59:58Apple M3 Max [...] I do push it

Reviewers have a hard time to fully load top Apple M chips. So can you please tell us your good application use of them?

Quotex86 is a dead end.

If we judge this propaganda by the tiny decrease of the percentage of Windows (x86, x64) and Linux combined, x86 / x64 will live for another 100 years.

I would LOVE to see nVidia put out a great ARM chip to be used for gaming, and hope they partner with Steam  (Valve) closely on this since both companies could benefit from alternatives to x86 Windows architecture. In addition to my AMD Alienware gaming system, I use a MacBook Pro that has an Apple M3 Max (ARM) chip and it is by far the fastest, most powerful CPU I've ever used AND uses surprisingly little electricity and runs almost silently. I don't game on my Mac, but I do push it, and I'm 100% convinced that a comparable CPU could run a gaming PC with a discrete GPU and the CPU would use WAY less resource than it currently needs to on x86 Windows architecture. Qualcomm is the only other company I think is really pushing great ARM chips, but nVidia is well positioned to be a strong player here, and Intel and AMD need to up their game in the ARM side of things. x86 is
a dead end.

Quote from: sirsquishy on November 03, 2024, 19:35:41Really just x86. Most things MS are Azure/Entra and cloud now. And any legacy app can be thrown behind VDI on...Azure/Entra VMs! Lol.

x86 is the real hold over. But is it really? Look at Nintendo's console, what we can do on iOS/Android, and the Apple M series. ARM has also been breaking into the datacenter for a long time with arrays of cores, forcing x86 to keep up (mostly by AMD, Intel is still a no show).

All software has to do is decide if ARM is worth the time/money and its all over for x86. Which would also force the hand of hold over companies stuck on 1990's platforms. if x86 goes poof, their end user experience goes with it. Forcing huge platform upgrades in order to roll over to ARM end user compute.

ARM vs x86 might be the best thing left for the industry from a technology perspective. x86 is pretty much at its peak now deep in a wide core race (look at AMD's 7000 vs 9000 debate).

If only IBM had picked a decent CPU in the beginning...

sigh

In manufacturing circles the 'holy grail' is having multiple vendors able to supply the same parts.

With x86 it is rather single/dual vendor with each of them 'mostly' compatible software wise however the hardware is grossly incompatible. Intel is infamous for 'reinventing' its platforms EVERY YEAR rendering ANY ability to leverage existing designs nary impossible.

'ARM' is a license that almost countless companies buy into and develop products with.

So in the end, get a OEM to present a 'platform' to a handful of chip developers and ask: "make an APU for this pinout configuration" and maybe, just maybe we might evolve to a market where multiple vendors are producing competitive APU's for the same platform.

:)

Quote from: GeorgeS on November 03, 2024, 17:27:32No surprise here.

While Qualcomm may of paved the way with their incompatible & under performing devices others have (Apple for example) successfully (more or less) switched over to ARM based systems from x86.

However it Apples case they may have a more robust Emulator AND much smaller number of developers to convince to compile native applications.

Adding extra players in the 'win-on-ARM' game only makes sense if there is a COMPELLING feature or reason to support the olatform other than 'it is there'.

End users and corporate $$$ can easily be better spent purchasing something powered by an AMD chip (or even Intel) to ensure the best compatibility.

Long awaited, it will be interesting to see what they come up with.

Keeping in mind there are TWO structures to break away from - both x86 AND Microsoft.

Really just x86. Most things MS are Azure/Entra and cloud now. And any legacy app can be thrown behind VDI on...Azure/Entra VMs! Lol.

x86 is the real hold over. But is it really? Look at Nintendo's console, what we can do on iOS/Android, and the Apple M series. ARM has also been breaking into the datacenter for a long time with arrays of cores, forcing x86 to keep up (mostly by AMD, Intel is still a no show).

All software has to do is decide if ARM is worth the time/money and its all over for x86. Which would also force the hand of hold over companies stuck on 1990's platforms. if x86 goes poof, their end user experience goes with it. Forcing huge platform upgrades in order to roll over to ARM end user compute.

ARM vs x86 might be the best thing left for the industry from a technology perspective. x86 is pretty much at its peak now deep in a wide core race (look at AMD's 7000 vs 9000 debate).

No surprise here.

While Qualcomm may of paved the way with their incompatible & under performing devices others have (Apple for example) successfully (more or less) switched over to ARM based systems from x86.

However it Apples case they may have a more robust Emulator AND much smaller number of developers to convince to compile native applications.

Adding extra players in the 'win-on-ARM' game only makes sense if there is a COMPELLING feature or reason to support the olatform other than 'it is there'.

End users and corporate $$$ can easily be better spent purchasing something powered by an AMD chip (or even Intel) to ensure the best compatibility.

Long awaited, it will be interesting to see what they come up with.

Keeping in mind there are TWO structures to break away from - both x86 AND Microsoft.

For real competition it would be right to deprive Intel/AMD of patents on x86, but the zombie company Intel is now being vigorously saved by non-market methods of the Western authorities, so depriving them of patents is out of the question. Forget about real market competition, it has long been absent in the West with an idiotic system of patent law made in favor of fat "whales" resting on the laurels of the past, and not nimble advanced startups.

NikoB.

This attempt is about the same level of value as Qualcomm's attempts. Without licenses for the x86 architecture, they will be forced to make something incompatible, which will be interesting only in narrow niches. And narrow niches - high factory cost and retail prices. A loss from the start. There is always a stumbling point where even sharks break their teeth...