[Libre-soc-bugs] [Bug 206] Implement branch prediction

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=206

--- Comment #30 from Jacob Lifshay <programmerjake at gmail.com> ---
I found a very interesting paper on the effects of modern branch predictors on
the performance of interpreters:
Branch Prediction and the Performance of Interpreters -
Don’t Trust Folklore
https://hal.inria.fr/hal-01100647/document

They tested a few Intel microarchitectures, as well as the combination of TAGE
and ITTAGE (which we very much should implement if we want decent cpu
performance).

Important takeaway: we *need* a good indirect-branch predictor (ITTAGE is a
good one). They can reduce branch mis-predictions to around 1-2 per thousand
instructions in stereotypical interpreter inner loops -- waay better than I
expected.

-- 
You are receiving this mail because:
You are on the CC list for the bug.