diff options
| author | Alasdair Armstrong <alasdair.armstrong@cl.cam.ac.uk> | 2019-01-22 18:50:25 +0000 |
|---|---|---|
| committer | Alasdair Armstrong <alasdair.armstrong@cl.cam.ac.uk> | 2019-01-22 18:50:25 +0000 |
| commit | 43e99a758f02f5c3a59732e658bab4f36e0a2400 (patch) | |
| tree | e4b82ebe67e6c2709fcdafbb51c515f93f127cfd /README.md | |
| parent | 96315109c65958c28246fb78eb55e6c5d6c0c296 (diff) | |
| download | sail-riscv-43e99a758f02f5c3a59732e658bab4f36e0a2400.zip sail-riscv-43e99a758f02f5c3a59732e658bab4f36e0a2400.tar.gz sail-riscv-43e99a758f02f5c3a59732e658bab4f36e0a2400.tar.bz2 | |
Mention that a faster model w/o tracing is available on the optimize
branch
This is the model for the current best performance numbers so we
should be clear that it is from this version. I want to merge these
changes back soon, but I need a nice way to enable/disable tracing
(and ensure zero-overhead when disabled) in a transparent way first.
Diffstat (limited to 'README.md')
| -rw-r--r-- | README.md | 5 |
1 files changed, 4 insertions, 1 deletions
@@ -115,7 +115,10 @@ the model should be run as: $ ./c_emulator/riscv_sim -t console.log -b spike.dtb bbl > execution-trace.log 2>&1 & $ tail -f console.log ``` -The `console.log` file contains the console boot messages. +The `console.log` file contains the console boot messages. For maximum +performance and benchmarking a model without any execution tracing is +available on the optimize branch (`git checkout optimize`) of this +repository. This currently requires the latest Sail built from source. Booting Linux with the OCaml backend ------------------------------------ |