|
|
@e2f601f
|
3 years |
Thierry Delisle |
Merge branch 'master' of plg.uwaterloo.ca:software/cfa/cfa-cc
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@8cd5434
|
3 years |
Thierry Delisle |
Added function to poll ready-queue length.
Not officially supported.
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@69914cbc
|
3 years |
mlbrooks |
Replacing "Mike's old linked list" with "Mike's new linked list," …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@1680072
|
3 years |
Thierry Delisle |
Changed ready_queue to use forloop instead of hardcoded unfolded loop of 2
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@f302d80
|
3 years |
Thierry Delisle |
Fix timestamp with new subqueue which was read after being cleared.
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@f55d54d
|
3 years |
Thierry Delisle |
Removed call to polymorphic 'min' which caused MASSIVE slowdown.
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@d3ba775
|
3 years |
Thierry Delisle |
More clean-up after new subqueue
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@eeb9f9f
|
3 years |
Thierry Delisle |
Forgot some changes for new sub-queue
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@f6fdfb14
|
3 years |
Thierry Delisle |
Removed old sub-queue
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@2b96031
|
3 years |
Thierry Delisle |
Added new subqueue implementation.
Seems faster will test on another …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@c993b15
|
3 years |
Thierry Delisle |
Changed RW lock to avoid hitting the global array on schedule.
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@b2fc7ad9
|
3 years |
Thierry Delisle |
Fixed missing return.
Was incorrectly removed in …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@fc59df78
|
3 years |
Thierry Delisle |
Split ready-queue routines in 3 instead of 2.
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@341aa39
|
3 years |
Thierry Delisle |
Fix bugs in workstealing edge cases:
- cutoff was wrong of all local …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@ddd473f
|
3 years |
Thierry Delisle |
Fix a copy/paste mistake and removed a warning in nodebug
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@d2fadeb
|
3 years |
Thierry Delisle |
Changed stats to make sense with relaxed fifo and work stealing
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@5f6a172
|
3 years |
Thierry Delisle |
Fix assertions on ready_queue with workstealling
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@431cd4f
|
3 years |
Thierry Delisle |
Added alternative to relaxed-fifo scheduler.
Disabled by default
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@9cc3a18
|
3 years |
Thierry Delisle |
Major clean-up before attempting to add new scheduler
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@e0d6748
|
3 years |
Thierry Delisle |
Removed unused functions
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@78d6c803
|
3 years |
Thierry Delisle |
Removed snzi from ready queue.
It hasn't been used in a while and I …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@a017ee7
|
3 years |
Thierry Delisle |
Ready-queue grow/shrink now reassigns the id of all processors.
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@a33c113
|
3 years |
Thierry Delisle |
Minor changes so using the global RWlock is more concise.
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@5cb51502
|
3 years |
Thierry Delisle |
Fix stats so they are correctly tallied when called from outside the …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@7a2972b9
|
3 years |
Thierry Delisle |
ready queue can now toggle between
- lock-based queue
- mpsc_queue …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@c426b03
|
3 years |
Thierry Delisle |
Minor clean-up
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@a344425
|
3 years |
Thierry Delisle |
Messed up the condition for external push, fixed now
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@fd1f65e
|
3 years |
Thierry Delisle |
Stats now keep track of external pushes separately
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@bd0bdd37
|
3 years |
Thierry Delisle |
Fix how bias is handled in the ready queue to be more consistent with …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@3143f28
|
3 years |
Thierry Delisle |
Fixed misleading stats counter
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@32c2c5e
|
4 years |
Thierry Delisle |
Removed unused variables
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@8fc652e0
|
4 years |
Thierry Delisle |
Change usage of TLS to more strongly segregate in kernel and out of …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@ceb7db8
|
4 years |
m3zulfiq |
Added alloc interface with backtick and removed polymorphic alloc …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@5fe7322
|
4 years |
Thierry Delisle |
Implemented reverse-rng strategy
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@772411a
|
4 years |
Thierry Delisle |
Moved bias to it's own function.
Fixed minor assertions triggering.
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@e67a82d
|
4 years |
pabuhr |
fix conflicts
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@fd9b524
|
4 years |
pabuhr |
change from asm volatile(pause) to Pause()
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@62502cc4
|
4 years |
Thierry Delisle |
Fixed deadlock where threads could acquire the central scheduler lock …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@343d10e
|
4 years |
Thierry Delisle |
Removed code broken in new-ast out of libcfa
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@1eb239e4
|
4 years |
Thierry Delisle |
Removed snzi and replaced it with a fast/slow path
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@39fc03e
|
4 years |
Thierry Delisle |
Fixed ready_queue working with 0/1 processors on cluster
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@320ec6fc
|
4 years |
Thierry Delisle |
Changed ready_queue_(grow/shrink) to take a target instead of going …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@28d73c1
|
4 years |
Thierry Delisle |
Moved processor registration to constructor
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@d72c074
|
4 years |
Thierry Delisle |
Added option to bias threads to queues, instead of processors to queues
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@52769ba
|
4 years |
Thierry Delisle |
Added statistics for local success when biased
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@13c5e19
|
4 years |
Thierry Delisle |
- Moved snzi and subqueues outside of ready_queue.cfa.
- Added …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@64a7146
|
4 years |
Thierry Delisle |
Fixed idle sleep to no-longer use a spinlock, broke registration and …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@04b5cef
|
4 years |
Thierry Delisle |
Added BIAS option to ready_queue
Added yield test for LibFibre?
Fixed …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@37ba662
|
4 years |
Thierry Delisle |
Minor improvements to alignments and memory layout
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@8834751
|
4 years |
Thierry Delisle |
Moved statistics to stats.cfa to combine ready Q stats and IO stats
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@9b1dcc2
|
4 years |
Thierry Delisle |
Changed scheduling API to adapt to non-Processors scheduling threads.
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@b388ee81
|
4 years |
Thierry Delisle |
Changed ready RW-Lock to be a single global lock instead of per …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@61d7bec
|
4 years |
Thierry Delisle |
Replaced the bitmask approached for the ready-queue with a SNZI
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@1b143de
|
4 years |
Thierry Delisle |
Several fixes to relaxed_ready queue
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@504a7dc
|
4 years |
Thierry Delisle |
Some fixes after the merge, compiles but still has livelocks
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@dca5802
|
5 years |
Thierry Delisle |
Started doing some of the x86 implementations and some changes after a …
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@c84b4be
|
5 years |
Thierry Delisle |
new ready queue seems to work but halting does not, had to be disabled
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@983edfd
|
5 years |
Thierry Delisle |
Ready queue seems to work but clearly shows quiescing is broken
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@b798713
|
5 years |
Thierry Delisle |
Working ready queue
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|
|
|
@7768b8d
|
5 years |
Thierry Delisle |
First step at adding the new ready queue to Cforall
ADTarm-ehast-experimentalenumforall-pointer-decayjacob/cs343-translationnew-astnew-ast-unique-exprpthread-emulationqualifiedEnum
|