Viewing profile — nsteel

19.2Tb is 2 x 9.6Tbps. Some (most) companies count Tx and Rx separately despite it making no sense in the context of serdes lanes. Stupid marketing in my opinion. I don't think the…

Isn't it 6x 200Gb octals? An octal being 8x 200Gb lanes. So 9.6Tbps?

Apparently it's Broadcom for 8t and Mediatek for 8i. https://wccftech.com/google-splits-tpuv8-strategy-two-chips-...

This link has more on the architecture: https://cloud.google.com/blog/products/compute/tpu-8t-and-tp...

It's HBM3e

It's similar to RAID schemes but instead of drive failure it's port unavailability. There's a reference at [1] or an FPGA-centric one at [2], but it applies to anywhere where dual/…

Also to cheaply (area) create multi-port RAMs.

And and Broadcom designs a huge part of the chip. They take Google's (mostly) logical design and providing everything TSMC need to physically make the chip (including imports g IP …

OK, but that's significantly slower and larger. It's a worse solution. Am I missing something?

I disagree. We've produced numerous complex chips with VHDL over the last 30 years. Most of the vendor models we have to integrate with are Verilog, so perhaps it is more popular, …

Except they don't use DDR5. LPDDR5 is always soldered. LPDDR5 requires short point-to-point connections to give you good SI at high speeds and low voltages. To get the same with DD…

> but it's always getting updated, I entirely disagree. Due to a combination of ISPs sticking with what they know and refusing to update (because of the huge time/cost in validatin…

I don't know if this is still the case, but back then the likes of Synopsys charged a lot of money for what was very limited controller functionality; you were stuck with their fru…

> the easiest part of the system to get changed... the core routers and associated infrastructure. Is that really the easy bit to change? ISPs spend years trialling new hardware an…

Implementing DDR3 training for our packet queuing chip (custom memory controller) was my first project at work. We had originally hoped to use the same training params for all part…

https://news.ycombinator.com/item?id=47256032 M1 Air single core is 2347 and multicore is 8342. Plus, the A18 pro chip in the Neo will likely perform better with the improved therm…

And that's why it's got to be a big company that takes this on, you need deep pockets to successfully sue a company like Intel. It's not realistic for most. Plus the huge opportuni…

Doesn't that M1 Air have a much, much slower CPU?

Worthless. Just looks how IFS worked out the previous two times they gave it a go. If you're not in the industry you may not even be aware it was a thing. And then not. Twice.

Agree entirely with your take. The packaging story is awesome, I wish there were more details on the stacking used on this one. But I am at a loss to how Intel are really going to …

> It’s a platform that sucks artists for everything they have, it actively prevents community building The API used to be great. It's slowly being crippled, presumably a PR/face-sa…

Japan is a terrible example for you, they are focused on ditching the US.

Hang on, they used a computer to help them create the post content?! Outrageous.

What about Etched? https://techfundingnews.com/nvidia-rival-ai-chip-maker-etche...

I think Bandcamp could do with a simple way to just pay artists money. There is that "name your price" thing but it's a bodge.