4 ms·

I would argue there were many good things but not well delivered. The Optane persistent memory should've been revolutionary for databases but Intel just put it

by alecco 2y ago

I would argue there were many good things but not well delivered. The Optane persistent memory should've been revolutionary for databases but Intel just put it out and expected people to do all the software. I'm seeing the same thing now with Intel QAT / IAA / DSA. Only niche software support. Only AWS seems to have it and those "bare metal" machines don't even have local NVMe. About 10 years ago Intel Research was publishing a lot of great research but no software for the users. Contrast it with Nvidia and their amazing software stack and support for their hardware.

gondo 2y ago
> Nvidia and their amazing software stack and support for their hardware. Linus seems to disagree https://m.youtube.com/watch?v=tQIdxbWhHSM https://m.youtube.com/watch?v=tQIdxbWhHSM
- mnau 2y ago
  His comment has nothing to do with quality of software or quality of support, but is about dealing with NVidia. Trying to work with NVidia (as a hardware manufacturer) must have been frustrating, but that has nothing to do with quality of the software. The video is 12 years old. A lot changed in the meantime. AMD has open source drivers and crashes often. NVidia has (or more precisely had) closed source drivers that nearly always work.
- alecco 2y ago
  That was about drivers and in 2012. At the time Linux was not interesting for them as clients. But now with AI, Nvidia has open source drivers. https://www.phoronix.com/news/NVIDIA-Transitions-OSS-KMD https://www.phoronix.com/news/NVIDIA-Transitions-OSS-KMD
  - akimbostrawman 2y ago
    >Nvidia has open source drivers. Kernel Module Driver which is most likely less than 5% of drivers.
- lmm 2y ago
  Torvalds wants open drivers, and NVidia doesn't do that. NVidia's drivers are better than their competitors by enough to make it worth buying NVidia even when their hardware is objectively worse, so much as I would prefer open-source in principle, I can understand why they don't want to give away the crown jewels. (I can't watch videos)
jiggawatts 2y ago
When I read the Intel QAT / IAA / DSA whitepaper I knew it was the beginning of the end for Intel. Every aspect of that document was just dripping in corporate dinosaur / MBA practices. For example, they include 4 cores of these accelerators in most of their Xeons, but soft fuse them off unless you buy a license. Nobody is going to buy that license. Okay, maybe one or two hyperscalers, but nobody else for certain. It's ultra-important with a feature like this to make it available to everybody, so that software is written to utilise it. This includes the starving university student contributing to Postgres, not just some big-enterprise customer that merely buys their software! They're doing the same stupid "gating" with AVX-512 as well, where it's physically included in desktop chips, but it is fused off so that server parts can be "differentiated". Meanwhile AMD just makes one compute tile that has a uniform programming API across both desktop and server chips. This means that geeks tuning their software to run on their own PCs are inadvertently optimising them for AMD's server chips as well! PS: Microsoft figured this out a while ago and they fixed some of their products like SQL Server. It now enables practically all features in all SKUs. Previously when only Enterprise Edition has certain programmability features nobody would use them because software vendors couldn't write software that customers couldn't install because they only had Standard Edition!
- alecco 2y ago
  Preach. I gave up hoping on Intel after getting burned. I feel sorry for the good talent there and I hope they find a home somewhere better.
  - shiroiushi 2y ago
    It sounds like most of the good talent has already left; the people still there are just there for the paycheck or resume padding and not for the long-term.
- singhrac 2y ago
  You see this again and again - when you gate technical features behind an enterprise barrier, the adoption rate is very low. AMD messed this up with CDNA and RDNA (now merging to UDNA), but Nvidia had it right from the start: CUDA from the bottom to the top. Intel started doing this kind of right recently with their Xe cores (ideally same programming model between their integrated GPUs and their datacenter HPC GPUs), but we’ll see how different the Xe-LPG and Xe-HPC end up being when Falcon Shores really comes out (I’m very worried about this description of it as a XPU, which seems like real confusion about what ML people want).

by alecco 2y ago

gondo 2y ago

> Nvidia and their amazing software stack and support for their hardware. Linus seems to disagree https://m.youtube.com/watch?v=tQIdxbWhHSM https://m.youtube.com/watch?v=tQIdxbWhHSM

mnau 2y ago
His comment has nothing to do with quality of software or quality of support, but is about dealing with NVidia. Trying to work with NVidia (as a hardware manufacturer) must have been frustrating, but that has nothing to do with quality of the software. The video is 12 years old. A lot changed in the meantime. AMD has open source drivers and crashes often. NVidia has (or more precisely had) closed source drivers that nearly always work.
alecco 2y ago
That was about drivers and in 2012. At the time Linux was not interesting for them as clients. But now with AI, Nvidia has open source drivers. https://www.phoronix.com/news/NVIDIA-Transitions-OSS-KMD https://www.phoronix.com/news/NVIDIA-Transitions-OSS-KMD
- akimbostrawman 2y ago
  >Nvidia has open source drivers. Kernel Module Driver which is most likely less than 5% of drivers.
lmm 2y ago
Torvalds wants open drivers, and NVidia doesn't do that. NVidia's drivers are better than their competitors by enough to make it worth buying NVidia even when their hardware is objectively worse, so much as I would prefer open-source in principle, I can understand why they don't want to give away the crown jewels. (I can't watch videos)

jiggawatts 2y ago

When I read the Intel QAT / IAA / DSA whitepaper I knew it was the beginning of the end for Intel. Every aspect of that document was just dripping in corporate dinosaur / MBA practices. For example, they include 4 cores of these accelerators in most of their Xeons, but soft fuse them off unless you buy a license. Nobody is going to buy that license. Okay, maybe one or two hyperscalers, but nobody else for certain. It's ultra-important with a feature like this to make it available to everybody, so that software is written to utilise it. This includes the starving university student contributing to Postgres, not just some big-enterprise customer that merely buys their software! They're doing the same stupid "gating" with AVX-512 as well, where it's physically included in desktop chips, but it is fused off so that server parts can be "differentiated". Meanwhile AMD just makes one compute tile that has a uniform programming API across both desktop and server chips. This means that geeks tuning their software to run on their own PCs are inadvertently optimising them for AMD's server chips as well! PS: Microsoft figured this out a while ago and they fixed some of their products like SQL Server. It now enables practically all features in all SKUs. Previously when only Enterprise Edition has certain programmability features nobody would use them because software vendors couldn't write software that customers couldn't install because they only had Standard Edition!

alecco 2y ago
Preach. I gave up hoping on Intel after getting burned. I feel sorry for the good talent there and I hope they find a home somewhere better.
- shiroiushi 2y ago
  It sounds like most of the good talent has already left; the people still there are just there for the paycheck or resume padding and not for the long-term.
singhrac 2y ago
You see this again and again - when you gate technical features behind an enterprise barrier, the adoption rate is very low. AMD messed this up with CDNA and RDNA (now merging to UDNA), but Nvidia had it right from the start: CUDA from the bottom to the top. Intel started doing this kind of right recently with their Xe cores (ideally same programming model between their integrated GPUs and their datacenter HPC GPUs), but we’ll see how different the Xe-LPG and Xe-HPC end up being when Falcon Shores really comes out (I’m very worried about this description of it as a XPU, which seems like real confusion about what ML people want).