-
Notifications
You must be signed in to change notification settings - Fork 572
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[BLIS] Update to 1.1 #10284
base: master
Are you sure you want to change the base?
[BLIS] Update to 1.1 #10284
Conversation
Co-authored-by: Mosè Giordano <[email protected]>
Well, this is interesting. Apparently what we have here as 1.0 from commit flame/blis@6d0ab74 is actually just a random commit from after 1.0 was released, since 1.0 is commit flame/blis@49af224, and so our current version is closer to 2.0 than 1.0 since they had started to make major architectural changes already (hence why the patches are failing). |
Yes, this was noted at the time in #8626 |
Yeah these patches are annoying to maintain - not hard but just annoying. Perhaps this release can even be skipped and the right thing is to wait for 2.0. |
CC @xrq-phys |
|
||
# Screen out A64FX sector cache. | ||
patch config/a64fx/bli_cntx_init_a64fx.c ${WORKSPACE}/srcdir/patches/a64fx_config_screen_sector_cache.patch | ||
fi | ||
#patch config/a64fx/bli_cntx_init_a64fx.c ${WORKSPACE}/srcdir/patches/a64fx_config_screen_sector_cache.patch |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
My understanding is that removing these patches is going to make performance worse
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yeah I was just seeing if the build goes through.
This is a bug fix release of various things since 1.0 last year. They already have a 2.0 RC out actually (which looks pretty cool...)
I have enabled aarch64 freebsd to see what happens. I haven't looked into how RISCV is implemented in the BLIS targets though, so I haven't added that arch.