Performance improvement when using "QEMU_FLATTEN" with softfloat type conversions
Bug #1892081 reported by
Ahmed Karaman
This bug affects 1 person
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| QEMU |
Fix Released
|
Undecided
|
Unassigned | ||
Bug Description
Attached below is a matrix multiplication program for double data
types. The program performs the casting operation "(double)rand()"
when generating random numbers.
This operation calls the integer to float softfloat conversion
function "int32_
Adding the "QEMU_FLATTEN" attribute to the function definition
decreases the instructions per call of the function by about 63%.
Attached are before and after performance screenshots from
KCachegrind.
Revision history for this message
| Ahmed Karaman (ahmedkrmn) wrote | #1 |
Revision history for this message
| Ahmed Karaman (ahmedkrmn) wrote | #2 |
Revision history for this message
| Ahmed Karaman (ahmedkrmn) wrote | #3 |
Revision history for this message
| Richard Henderson (rth) wrote | #4 |
| Changed in qemu: | |
| status: | New → In Progress |
| assignee: | nobody → Richard Henderson (rth) |
Revision history for this message
| Thomas Huth (th-huth) wrote Moved bug report | #5 |
| Changed in qemu: | |
| assignee: | Richard Henderson (rth) → nobody |
| status: | In Progress → Expired |
| Changed in qemu: | |
| status: | Expired → Fix Committed |
| Changed in qemu: | |
| status: | Fix Committed → Fix Released |
To post a comment you must log in.
Confirmed, although "65% decrease" is on 0.44% of the total
execution for this test case, so the decrease isn't actually
noticeable.
Nevertheless, it's a simple enough change.