git.annexia.org
/
riscv-talks.git
/ blobdiff
commit
grep
author
committer
pickaxe
?
search:
re
summary
|
shortlog
|
log
|
commit
|
commitdiff
|
tree
raw
|
inline
| side by side
2023-koji: Add scaling notes
[riscv-talks.git]
/
2023-hardware
/
notes.txt
diff --git
a/2023-hardware/notes.txt
b/2023-hardware/notes.txt
index
c13f6d7
..
62bcdb3
100644
(file)
--- a/
2023-hardware/notes.txt
+++ b/
2023-hardware/notes.txt
@@
-187,3
+187,8
@@
A few tips to make TCG run (a bit) fast(er):
- Software TLB
- Fast vs slow jumps
- Software TLB
- Fast vs slow jumps
+
+ - Performance scales almost linearly with # vCPUs
+
+ * I compared 4 vCPU vs 16 vCPU virtual machine, and parallelized
+ build speed inside the guest increased by 3.3x, which is almost linear