From d29b16e5b4e5fab1224344e5537fbdf7fffbdd4a Mon Sep 17 00:00:00 2001 From: George Hotz Date: Wed, 16 Jun 2021 11:47:57 -0700 Subject: [PATCH] more business notes --- fpga/README | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/fpga/README b/fpga/README index 333ccd115e..994ffac283 100644 --- a/fpga/README +++ b/fpga/README @@ -15,6 +15,10 @@ Small Board (Arty A7 100T) * 4x4x4 matmul = 64 mults, perhaps 8x8x8 matmul = 512 mults * 6.4 GFLOPS @ 50 mhz +* Forward/backward pass of ResNet-50, EfficientNet-B2, and BERT-large in the simulator +* Train MNIST models on the real hardware +* After we've trained MNIST here, buy the big board and a Linux computer for home + Big Board (Alveo U250) ===== * Support DMA over PCI-E. 16 GB/s @@ -24,6 +28,12 @@ Big Board (Alveo U250) * 16x16x16 matmul = 4096 mults, perhaps 32x32x32 matmul = 32768 mults * 4 TFLOPS @ 500 mhz +* Bring up in one Z840 with one card +* Train (with tinygrad) ResNet-50, EfficientNet-B2, and BERT-large +* Now we buy a machine with 8x cards +* Write 8x multicard training, place on https://mlcommons.org/en/training-normal-07/ +* Now it's funding/kickstarter time, based on our MLPerf results on the Alveos and Cherry Two sim + Cherry Two (12nm tapeout) ===== * Support DMA over PCI-E. 16 GB/s @@ -34,6 +44,10 @@ Cherry Two (12nm tapeout) * Target 75W, even if underclocked. One slot, no external power. * This card should be on par with a 3090 and sell for $1000 +* Write PyTorch port to support same training while waiting for tapeout +* If we are here, we are winning the AI chip market +* Tile the core and go to a smaller process node + Cherry Three (5nm tapeout) ===== * Support DMA over PCI-E 4.0. 32 GB/s