A question about "Software rendering for gnome-shell"
mwesten
mwesten at verizon.net
Thu Mar 1 18:01:10 UTC 2012
On 02/23/2012 09:54 AM, Adam Jackson wrote:
> In the non-steady-state, though, the current implementation is known
> to be incredibly memcpy-heavy. Fixes coming soon, beta should be
> much better I hope. If you want to collect some data about where CPU
> time is being spent, 'perf record' against the X server or
> gnome-shell (with debuginfo installed) and then 'perf report' should
> be enlightening. I suspect you'll find the vast majority of the
> time spent in memcpy in one form or another (pixman or fb blit,
> copying data into and out of the kernel across the unix socket,
> etc).
>
> - ajax
>
I set this up on the Alpha release and the results are below. I'm
assuming there's something wrong with that top entry, but I don't know
how to fix it.
-Mike
Events: 50K cycles
55.99% gnome-shell perf-6425.map [.] 0x19b99e
4.91% gnome-shell swrast_dri.so [.]
lp_tile_b8g8r8x8_unorm_swizzle_4ub
3.50% gnome-shell swrast_dri.so [.]
lp_tile_b8g8r8x8_unorm_unswizzle_4ub
2.60% gnome-shell libX11.so.6.3.0 [.] _XPutPixel32
2.57% gnome-shell libX11.so.6.3.0 [.] _XGetPixel32
2.40% gnome-shell [kernel.kallsyms] [k]
__copy_from_user_ll
1.97% gnome-shell swrast_dri.so [.]
lp_rast_clear_zstencil
1.62% gnome-shell [kernel.kallsyms] [k]
get_page_from_freelist
1.49% gnome-shell [kernel.kallsyms] [k]
__lock_acquire
1.03% gnome-shell libc-2.15.so [.]
__memmove_ia32
0.97% gnome-shell [kernel.kallsyms] [k]
native_sched_clock
0.85% gnome-shell libX11.so.6.3.0 [.] _XSetImage
0.69% gnome-shell libc-2.15.so [.] __memcpy_ia32
0.66% gnome-shell [kernel.kallsyms] [k]
sched_clock_local
0.64% gnome-shell [kernel.kallsyms] [k]
__kernel_text_address
0.63% gnome-shell libX11.so.6.3.0 [.]
__x86.get_pc_thunk.bx
0.52% gnome-shell swrast_dri.so [.]
lp_rast_shade_tile
More information about the test
mailing list