Or is there a reason that this won't work (or at least won't be worth it) for ANN structures?
If all the weights and activations fit in on-chip memory, then you can do calculations at close to 100% efficiency. If you want to simulate a 20k times larger network, you also need to transfer the 4 Billion parameters per iteration, which would take a singificantly longer time. In other words, you would be seriously bottlenecked.
In silicon we separate out the storage from the compute. Generally we have much more storage than we have compute, and we route the data to the compute units as we need computation. Therefore ALUs have aggregate throughput of only a small fraction of the model capacity per cycle; at 400,000 cores with, say, 8 16-bit operations per cycle each, that is ‘only’ ~10^6 operations per cycle.
A better way to get a measure of computational throughput is to compare those values. Numbers are hard to find, but 10^10 synaptic events/s for a bee and 10^15 edge calculations/s for Cerebras seems the right sort of ballpark. That's a difference of 10^5, a little less than your estimate of 10^7.
That means we'll need 200 such chips.
Why is the wafer so expensive ? Buying a manufactured wafer from TSMC ,16nm is I think is $4k-5k
And there are still a few nodes left. While cost per gate don't go down much, it could solve some of the external bandwidth issues.
But all those things are really expensive and very risky. Who knows...
When you have only a few giant chips per wafer, each error becomes devastating, taking out a large fraction of the wafer. Errors are so common I wouldnt be surprised if it took them 5-10 wafers per working chip, especially if they dont do any binning or they didnt design the chip to fuse off bad sections like Intel CPUs do (a single i3/i5/i7/i9 line is usually the same exact chip, with cores that have too many manufacturing errors fused off so perfect parts become i9s, less perfect parts become i7s, etc).
In the case of Cerebras, they have disable the bad blocks to get chips that actually work.
So, someone like Cerebras has to both make as little critical as possible and buy much more expensive wafers with lower defect rates and still get only moderate yields.
How do you do that ? where do i read more about this ?
Better wafers cost more.
That probably means their wafer yield is 100%.
As for wafer prices: https://anysilicon.com/major-pure-play-foundries-revenue-per...
It's $6k for all wafer smaller than 20nm, which means 16nm is somewhat cheaper than that.
Not trying to crap on the tech, but it's not so simple as we can merely go to the fab and get small quantities of full custom, 100% yielding chips, for 6k ea.
That doesn't make any sense because errors are an inherent part of the manufacturing process. If errors in the manufacturing process make 1% of the area unavailable that means perfect yields could at most be 99%.
CEO implies in the FYI podcast that it can handle model sizes of a max of ~4 billion parameters per wafer. Respectable but not as large as I assumed would be possible when I first read of the scale of the chip.
CEO claims model parallelism will actually work well with these devices. Would be intriguing to know the limits of this.
Based on the cooling requirements, at least a GHz clock speed appears likely. If we take the heretical position that 1 parameter is >= 1 synapse, 4 billion parameters is about the size of a bee brain. 1GHz is about 5 million times faster than a bee brain.
It would take about 20000 such chips to simulate a human-sized network. Not economical and presumably the model parallelism would break long before this. But it is interesting to note that under the not implausible 1 parameter is >= 1 synapse assumption, we are only a few orders of magnitude away from human-sized networks training 5 million times real-time.