You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Install the [NVIDIA Container Toolkit](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html) if you are using NVIDIA GPUs.
**Note:** When benchmarking, make sure that you run 1 round of Fibonacci (or some other small program) to download the R0 docker image to make sure it doesn't contribute to benchmarking time.
51
40
52
41
**Note:** On Ubuntu 22.04, you might need to install libssl1.0 to use the Risc0 toolchain. You can follow these instructions (from [here](https://stackoverflow.com/questions/72133316/libssl-so-1-1-cannot-open-shared-object-file-no-such-file-or-directory/73604364#73604364)).
53
42
54
-
## Machine Requirements
55
-
56
-
We ran our benchmarks on a AWS r7i.16xlarge machine with 64 vCPUs and 512 GB of memory. Smaller programs require significantly less memory. Note that the machine has to support AVX-512 instructions, otherwise the command in `eval.sh` will not run. If your machine doesn't support AVX-512, then modify this command to remove those flags.
57
-
58
43
## Usage
59
44
60
45
You can conduct a sweep of the benchmarks by running either of the following commands:
61
46
62
-
```
47
+
```sh
63
48
python3 sweep.py
64
49
```
65
50
66
-
or running the following command for the JOLT sweep:
0 commit comments