101–109 of 109 posts
Earlier quoted context omitted.
A 9700X is twice the performance of a 9900K and M5 Max is almost 3X the performance. The megahertz myth is a myth.
I replied to the sibling comment: I was making simplifying assumptions for two specific use cases and naively treated physical cores and clock rate as my variables.
Earlier quoted context omitted.
Even Apple hardware looks inexpensive compared to Nvidia's huge premium. And never mind the order backlog. x86 and Apple already sell CPUs with integrated memory and high bandwidth interconnects. And I bet eventually Intel's beancounter board will wake up and allow engineering to make one, too. But competition is good for the market.
Apple went from a high-end PC to a low-end AI provider due to blocking Nvidia on their platform.
They also dropped the ball on the data center, having left OS X Server behind.
Those markets are now served by Windows or Linux based configurations.
Does this mean their gaming GPUs are becoming less in demand, and therefore cheaper/more available again?
Absolutely not, unfortunately. The problem is not that gaming GPUs are in demand, it’s that selling silicon to AI center buildouts is so absurdly profitable right now they just aren’t making many gaming GPUs. If you can only get so many mm^2 of dies from TSMC, might as well make 50x selling to AI providers.
At least there are a few cool ones about programming CUDA directly in Python.
Earlier quoted context omitted.
It's a tech demonstrator for a company that turns models into custom silicon for fast inference. In this case llama3.1-8b https://taalas.com/products/
Is this an ASIC? Or FPGA? Or something even more exotic? I’m guessing it’s some form of ASIC because I can’t imagine crafting the logic of Llama on silicon is a very quick or easy job. Not that doing it on an ASIC is a piece of cake either.
Anyways, I found this article discussing it a bit more: https://www.eetimes.com/taalas-specializes-to-extremes-for-e...
"Taalas is borrowing some ideas from the structured ASICs of the early 2000s to make its hardwired model-specific chips. Structured ASICs used gate arrays and hardened IP blocks, changing only the interconnect layers to adapt the chip to a specific workload. At the time, this was seen as a more cost-effective alternative to a full-custom ASIC that was more performant than an FPGA."
"Taalas changes only two masks to customize a chip for a specific model, but the two masks can change both model weights and dataflow through the chip. On the HC1, the model and its weights are stored on the chip using a mask-ROM-based recall fabric paired with a (programmable) SRAM, which can be used to hold fine-tuned weights and/or the KV cache. Future generations of chips may split the SRAM onto a separate chip, meaning they could be denser than the HC1."
It is a 88-core ARM v9 chip, for somewhat more detailed spec.
Vera does what NVIDIA calls Spatial Multithreading, "physically partitioning each core’s resources rather than time slicing them, allowing the system to optimize for performance or density at runtime." A kind of static hyperthreading; you get two threads per core. It's somewhat different from how x86 chips do simultaneous multithreading (SMT),
In operating systems timeslicing means giving a quantum of execution time to each process, and context switching between processes. Not normally a term used in computer architecture but possibly the characterisation would fit a barrer processor rather than SMT.
Earlier quoted context omitted.
Absolutely not, unfortunately. The problem is not that gaming GPUs are in demand, it’s that selling silicon to AI center buildouts is so absurdly profitable right now they just aren’t making many gaming GPUs. If you can only get so many mm^2 of dies from TSMC, might as well make 50x selling to AI providers.
Check the GTC 2026 agenda, there are hardly any graphics programming talks. At least there are a few cool ones about programming CUDA directly in Python.
Other than Hyperscaler ARM has yet to enter the server market and it might well be Nvidia that makes a different.
Both still run web tech in a wrapper, just with different performance characteristics. The local-first vs cloud distinction is more fundamental, especially for tools that interact with platforms like LinkedIn.
When I built ZenMode, the core insight was that LinkedIn can easily detect automation coming from AWS/datacenter IPs, but when your desktop app uses your actual Chrome browser and home IP, it's indistinguishable from manual usage.
That's why we went with an Electron/Puppeteer architecture running locally rather than yet another cloud service. Check it out at https//zen-mode.io if you're curious about the local execution model.