4 ms·
Every time I see an FPGA article, I feel a little sad that Tabula[1] didn't make it -- 1.6Ghz clock and reprogrammable on the fly. RIP. 1. https://en.wikipedia
by thenobsta 3y ago
Every time I see an FPGA article, I feel a little sad that Tabula[1] didn't make it -- 1.6Ghz clock and reprogrammable on the fly. RIP.
1. https://en.wikipedia.org/wiki/Tabula,_Inc https://en.wikipedia.org/wiki/Tabula,_Inc.
- snvzz 3y agoNote the link is broken becuase the dot needs to be within the URL.
- hutzlibu 3y agoWhat happened with it? Wikipedia is very thin on the details. Did it simply not work out in mass production, or was it not considered worth it anymore as CPU and GPUs advanced?
- FullyFunctional 3y agoMy moles claim they simply couldn't get the software to work.
- bcrl 3y agoMy understanding is that the hardware turned out to use way more power than expected. The whole premise of the design was that the extra registers between logic stages were "free". In the real world, pipelining has a significant power cost.
- FullyFunctional 3y agoAre you talking about Achronix? Because they certainly did tell me that power was the reason they bailed on their original approach. Space-Time ran hot, but I don’t see why they would be using more flip-flops than with a conventional FPGA. Also my sources were pretty clear on the cause of their failure.
- FullyFunctional 3y agoI was a huge fan of the concept (time slicing LUTs). The name "Space-Time" actually made sense. Unfortunately, mapping designs to regular FPGA fabric is already hard and it apparently was too difficult for the Space-Time fabric and thus they folded. There are a lot of open FPGA efforts around. I hope someone else gives the idea another shot. While we are on exotic approaches, Achronix' first generation was using an asynchronous (1 GHz) fabric. I regret I didn't get a chance to play with that before the pivoted to a conventional fabric.