I don't understand on one of the later graphs the core to core latency for strix halo goes out to 32 cores but he says only has 16 cores?
AMD's cores have SMT, allowing them to run two threads at a time and appear to the OS and its scheduler as two logical cores despite being implemented as a single physical core.
What pattern in the data shows that's what's being measured? I would expect to see basically 0 latency between adjacent "cores" then since L1 is shared per thread?
Co resident threads might not get any speed up here since coherency instructions are functionally operations on the L2 cache.
In Europe these cost 5k EUR. I guess I'm not buying computer ever again and hopefully the 10 years old ones I have will never die.
DGX Spark vs. Strix Halo vs. M4 Max is hotly debated. You can find plenty of HN discussions and YouTube videos about it.
What’s stopping Mediatek from putting some of those cores in a laptop/desktop CPU package? Is it the infrastructure around it? OEM support? All of the above?
Most likely, it's just a matter of time and the NVIDIA DGX Spark is partly serving as a pipe cleaner product. They clearly need to work on power management, and solid Windows support may be more work than the Linux support they've shipped so far.
Mediatek-based laptops (other than the existing Chromebooks using what's more or less Mediatek phone chips) are one of the big things to keep an eye out for at CES next week. They have a solid market opportunity to provide an alternative ARM solution to compete against Qualcomm's Snapdragon X Elite and upcoming X2 Elite: having a NVIDIA GPU would give Mediatek a huge advantage over one of Qualcomm's most pronounced weaknesses, and the CPU cores Mediatek is using are probably "good enough" for a GPU-focused system (mobile AI workstation or low-power gaming laptop).
Kind of an odd choice to have clusters that are so different
I don't understand on one of the later graphs the core to core latency for strix halo goes out to 32 cores but he says only has 16 cores?
AMD's cores have SMT, allowing them to run two threads at a time and appear to the OS and its scheduler as two logical cores despite being implemented as a single physical core.
What pattern in the data shows that's what's being measured? I would expect to see basically 0 latency between adjacent "cores" then since L1 is shared per thread?
Co resident threads might not get any speed up here since coherency instructions are functionally operations on the L2 cache.
I assume that the author here is testing against one of these boxes, right? https://marketplace.nvidia.com/en-us/enterprise/personal-ai-...
Are these considered a good deal at $3-4k? What's the software support like on them? I've got 2x 3090s and I'm curious how this compares.
> https://marketplace.nvidia.com/en-us/enterprise/personal-ai-...
In Europe these cost 5k EUR. I guess I'm not buying computer ever again and hopefully the 10 years old ones I have will never die.
DGX Spark vs. Strix Halo vs. M4 Max is hotly debated. You can find plenty of HN discussions and YouTube videos about it.
What’s stopping Mediatek from putting some of those cores in a laptop/desktop CPU package? Is it the infrastructure around it? OEM support? All of the above?
Most likely, it's just a matter of time and the NVIDIA DGX Spark is partly serving as a pipe cleaner product. They clearly need to work on power management, and solid Windows support may be more work than the Linux support they've shipped so far.
Mediatek-based laptops (other than the existing Chromebooks using what's more or less Mediatek phone chips) are one of the big things to keep an eye out for at CES next week. They have a solid market opportunity to provide an alternative ARM solution to compete against Qualcomm's Snapdragon X Elite and upcoming X2 Elite: having a NVIDIA GPU would give Mediatek a huge advantage over one of Qualcomm's most pronounced weaknesses, and the CPU cores Mediatek is using are probably "good enough" for a GPU-focused system (mobile AI workstation or low-power gaming laptop).
Kind of an odd choice to have clusters that are so different