stefanp (Stefan Pintilie)
User

Projects

User does not belong to any projects.

User Details

User Since
May 29 2017, 8:02 AM (54 w, 6 d)

Recent Activity

Fri, Jun 15

stefanp added a dependency for D48218: [Power9] Add remaining __flaot128 builtin support for FMA round to odd: D47550: [Power9] Add __float128 builtins for Round To Odd.
Fri, Jun 15, 7:00 AM
stefanp added a dependent revision for D47550: [Power9] Add __float128 builtins for Round To Odd: D48218: [Power9] Add remaining __flaot128 builtin support for FMA round to odd.
Fri, Jun 15, 7:00 AM
stefanp created D48218: [Power9] Add remaining __flaot128 builtin support for FMA round to odd.
Fri, Jun 15, 7:00 AM

Thu, Jun 14

stefanp added a dependency for D48185: [Power9] [LLVM] Add __float128 exponent GET and SET builtins: D47550: [Power9] Add __float128 builtins for Round To Odd.
Thu, Jun 14, 12:02 PM
stefanp added a dependent revision for D47550: [Power9] Add __float128 builtins for Round To Odd: D48185: [Power9] [LLVM] Add __float128 exponent GET and SET builtins.
Thu, Jun 14, 12:02 PM
stefanp added a dependency for D48184: [Power9] [CLANG] Add __float128 exponent GET and SET builtins: D47550: [Power9] Add __float128 builtins for Round To Odd.
Thu, Jun 14, 12:01 PM
stefanp added a dependent revision for D47550: [Power9] Add __float128 builtins for Round To Odd: D48184: [Power9] [CLANG] Add __float128 exponent GET and SET builtins.
Thu, Jun 14, 12:01 PM
stefanp created D48185: [Power9] [LLVM] Add __float128 exponent GET and SET builtins.
Thu, Jun 14, 11:58 AM
stefanp created D48184: [Power9] [CLANG] Add __float128 exponent GET and SET builtins.
Thu, Jun 14, 11:56 AM

Wed, Jun 13

stefanp committed rL334613: [PowerPC] The __float128 type should only be available on Power9.
[PowerPC] The __float128 type should only be available on Power9
Wed, Jun 13, 9:09 AM
stefanp committed rC334613: [PowerPC] The __float128 type should only be available on Power9.
[PowerPC] The __float128 type should only be available on Power9
Wed, Jun 13, 9:09 AM
stefanp closed D48088: [PowerPC] The __float128 type should only be available on Power9.
Wed, Jun 13, 9:09 AM

Tue, Jun 12

stefanp created D48088: [PowerPC] The __float128 type should only be available on Power9.
Tue, Jun 12, 12:06 PM

Mon, Jun 11

stefanp updated the diff for D47548: [Power9] Add __float128 builtins for Round To Odd.

Updated the test case to use clang cc1 instead of just clang.

Mon, Jun 11, 8:40 AM

Thu, May 31

stefanp added a comment to D47550: [Power9] Add __float128 builtins for Round To Odd.

@lattner
Added Chris Lattner as a reviewer in case I did something silly in the target indep code.

Thu, May 31, 12:14 PM
stefanp added a reviewer for D47550: [Power9] Add __float128 builtins for Round To Odd: lattner.
Thu, May 31, 12:13 PM
stefanp updated the diff for D47550: [Power9] Add __float128 builtins for Round To Odd.

Fixed some of the style issues.

Thu, May 31, 12:12 PM
stefanp added a dependent revision for D47550: [Power9] Add __float128 builtins for Round To Odd: D47548: [Power9] Add __float128 builtins for Round To Odd.
Thu, May 31, 11:36 AM
stefanp added a dependency for D47548: [Power9] Add __float128 builtins for Round To Odd: D47550: [Power9] Add __float128 builtins for Round To Odd.
Thu, May 31, 11:36 AM
stefanp updated the summary of D47550: [Power9] Add __float128 builtins for Round To Odd.
Thu, May 31, 11:35 AM

Wed, May 30

stefanp created D47550: [Power9] Add __float128 builtins for Round To Odd.
Wed, May 30, 11:51 AM
stefanp created D47548: [Power9] Add __float128 builtins for Round To Odd.
Wed, May 30, 11:49 AM

Tue, May 22

stefanp added a comment to D45265: [PowerPC] Replace the Post RA List Scheduler with the Machine Scheduler.

Hi Eric,

Tue, May 22, 10:55 AM

May 17 2018

stefanp updated the diff for D45266: [PowerPC] Add a Memory Latency Mutation to the scheduler.

Modifications based on review comments.

May 17 2018, 10:57 AM

May 16 2018

stefanp requested review of D45265: [PowerPC] Replace the Post RA List Scheduler with the Machine Scheduler.
May 16 2018, 8:26 AM
stefanp added a comment to D45265: [PowerPC] Replace the Post RA List Scheduler with the Machine Scheduler.

Ran a series of tests on this patch.
Overall the results are positive. I'm going to list the degradations.

May 16 2018, 8:25 AM

May 3 2018

stefanp updated the diff for D42590: [PowerPC] Try to move the stack pointer update instruction later in the prologue and earlier in the epilogue (Version 2).

Address comments from previous review.

May 3 2018, 11:12 AM

Apr 20 2018

stefanp added inline comments to D39785: [PowerPC] Remove redundant register copies in late peephole.
Apr 20 2018, 2:01 PM

Apr 16 2018

stefanp updated the diff for D42590: [PowerPC] Try to move the stack pointer update instruction later in the prologue and earlier in the epilogue (Version 2).

I've updated the negative frame indices to use isFixedObjectIndex which is cleaner.

Apr 16 2018, 10:43 AM
stefanp added inline comments to D42590: [PowerPC] Try to move the stack pointer update instruction later in the prologue and earlier in the epilogue (Version 2).
Apr 16 2018, 10:41 AM
stefanp closed D44903: [LatencyPriorityQueue] The LatencyPriorityQueue class is missing the implementation for the dump function.

Committed in rL330117.
Closing this item.

Apr 16 2018, 4:04 AM
stefanp committed rL330117: [LatencyPriorityQueue] The LatencyPriorityQueue class is missing the….
[LatencyPriorityQueue] The LatencyPriorityQueue class is missing the…
Apr 16 2018, 3:26 AM

Apr 13 2018

stefanp committed rL330060: [Power9] Add the TLS store instructions to the Power 9 model.
[Power9] Add the TLS store instructions to the Power 9 model
Apr 13 2018, 12:54 PM

Apr 4 2018

stefanp created D45266: [PowerPC] Add a Memory Latency Mutation to the scheduler.
Apr 4 2018, 9:45 AM
stefanp added a dependent revision for D45265: [PowerPC] Replace the Post RA List Scheduler with the Machine Scheduler: D45266: [PowerPC] Add a Memory Latency Mutation to the scheduler.
Apr 4 2018, 9:45 AM
stefanp created D45265: [PowerPC] Replace the Post RA List Scheduler with the Machine Scheduler.
Apr 4 2018, 9:30 AM

Mar 28 2018

stefanp updated the diff for D42590: [PowerPC] Try to move the stack pointer update instruction later in the prologue and earlier in the epilogue (Version 2).

After some code cleanup that was required for requiresFrameIndexScavenging here is the new version of this patch.
It should take into consideration all of the reviewer's comments.

Mar 28 2018, 11:11 AM
stefanp added a comment to D44903: [LatencyPriorityQueue] The LatencyPriorityQueue class is missing the implementation for the dump function.

@echristo
Hi Eric,
What is your opinion on this? Should I be adding the #if .. and #endif? Does it matter?

Mar 28 2018, 8:08 AM

Mar 27 2018

stefanp committed rL328647: [Power9] Fix the resource list for the COPY instruction..
[Power9] Fix the resource list for the COPY instruction.
Mar 27 2018, 10:54 AM

Mar 26 2018

stefanp created D44903: [LatencyPriorityQueue] The LatencyPriorityQueue class is missing the implementation for the dump function.
Mar 26 2018, 11:29 AM
stefanp committed rL328556: [PowerPC] Infrastructure work. Implement getting the opcode for a spill in one….
[PowerPC] Infrastructure work. Implement getting the opcode for a spill in one…
Mar 26 2018, 10:42 AM
stefanp closed D43086: [PowerPC] Infrastructure work. Implement getting the opcode for a spill in one place..
Mar 26 2018, 10:42 AM
stefanp added inline comments to D43086: [PowerPC] Infrastructure work. Implement getting the opcode for a spill in one place..
Mar 26 2018, 10:13 AM

Mar 20 2018

stefanp updated the diff for D43086: [PowerPC] Infrastructure work. Implement getting the opcode for a spill in one place..

Implemented the cleanup for the loads as well.
Addressed comments from the reviewer.

Mar 20 2018, 6:28 PM

Mar 19 2018

stefanp updated the diff for D43086: [PowerPC] Infrastructure work. Implement getting the opcode for a spill in one place..

Added the XFormMemOp flag to the td files for the instructions the require it. This flag indicates that this instructions is a memory instruction and that it is an X-Form instruction.
Used a script to compare the modifications against the ISA to make sure that the correct instructions have the flag set.

Mar 19 2018, 12:24 PM

Mar 9 2018

stefanp committed rL327174: [Power9] Code Cleaup and adding Comments for Power 9 Scheduler.
[Power9] Code Cleaup and adding Comments for Power 9 Scheduler
Mar 9 2018, 1:13 PM
stefanp added a comment to D40855: [PowerPC] LSR tunings for PowerPC.

I've reverted the change in rL327143.

Mar 9 2018, 9:02 AM
stefanp committed rL327143: Revert "[PowerPC] LSR tunings for PowerPC".
Revert "[PowerPC] LSR tunings for PowerPC"
Mar 9 2018, 8:14 AM
stefanp committed rL327142: Revert "[PowerPC] Move test to correct location.".
Revert "[PowerPC] Move test to correct location."
Mar 9 2018, 8:14 AM
stefanp added a comment to D40855: [PowerPC] LSR tunings for PowerPC.

@echristo
Hi Eric,
Would you like me to pull this change out and fix the bug before I put it back in?

Mar 9 2018, 6:37 AM

Mar 8 2018

stefanp added a comment to D43235: [SchedModel] Complete models shouldn't match against itineraries when they don't use them (PR35639) (WIP).

With the patch rL327021 we should now be able to mark Power 9 as complete.

Mar 8 2018, 9:46 AM
stefanp committed rL327021: [Power9] Add more missing instructions to the Power 9 scheduler.
[Power9] Add more missing instructions to the Power 9 scheduler
Mar 8 2018, 8:27 AM

Mar 7 2018

stefanp committed rL326923: [PowerPC] Move test to correct location..
[PowerPC] Move test to correct location.
Mar 7 2018, 10:30 AM
stefanp committed rL326906: [PowerPC] LSR tunings for PowerPC.
[PowerPC] LSR tunings for PowerPC
Mar 7 2018, 8:57 AM
stefanp closed D40855: [PowerPC] LSR tunings for PowerPC.
Mar 7 2018, 8:56 AM

Mar 5 2018

stefanp committed rL326701: [Power9] Add more missing instructions to the Power 9 scheduler.
[Power9] Add more missing instructions to the Power 9 scheduler
Mar 5 2018, 6:38 AM

Mar 2 2018

stefanp committed rL326578: [Power9] Add missing instructions to the Power 9 scheduler.
[Power9] Add missing instructions to the Power 9 scheduler
Mar 2 2018, 6:44 AM

Mar 1 2018

stefanp committed rL326447: [Power9] Add missing instructions to the Power 9 scheduler.
[Power9] Add missing instructions to the Power 9 scheduler
Mar 1 2018, 8:18 AM
stefanp closed D43899: [Power9] Add more missing instructions to the Power 9 scheduler.
Mar 1 2018, 8:18 AM
stefanp added a comment to D43899: [Power9] Add more missing instructions to the Power 9 scheduler.

Thank you for the very quick review! :)
I'll add more comments to the files as I work on them over the next couple of days.

Mar 1 2018, 3:39 AM

Feb 28 2018

stefanp created D43899: [Power9] Add more missing instructions to the Power 9 scheduler.
Feb 28 2018, 1:14 PM

Feb 23 2018

stefanp committed rL325956: [Power9] Add missing instructions to the Power 9 scheduler.
[Power9] Add missing instructions to the Power 9 scheduler
Feb 23 2018, 12:39 PM
stefanp closed D43635: [Power9] Add missing instructions to the Power 9 scheduler.
Feb 23 2018, 12:39 PM
stefanp committed rL325918: [PowerPC] Code cleanup. Remove instructions that were withdrawn from Power 9..
[PowerPC] Code cleanup. Remove instructions that were withdrawn from Power 9.
Feb 23 2018, 7:58 AM
stefanp closed D43641: [PowerPC] Code cleanup. Remove instructions that were withdrawn from Power 9..
Feb 23 2018, 7:58 AM
stefanp added a comment to D43635: [Power9] Add missing instructions to the Power 9 scheduler.

I also learned that we accept regexes here when I started doing this a couple of days ago. :)

Feb 23 2018, 6:49 AM

Feb 22 2018

stefanp created D43641: [PowerPC] Code cleanup. Remove instructions that were withdrawn from Power 9..
Feb 22 2018, 12:19 PM
stefanp created D43635: [Power9] Add missing instructions to the Power 9 scheduler.
Feb 22 2018, 10:14 AM

Feb 12 2018

stefanp updated the diff for D43086: [PowerPC] Infrastructure work. Implement getting the opcode for a spill in one place..

Addressed comments from reviewer and removed work that was not really related to this patch.
Will split out the work that was not really related into a separate patch.

Feb 12 2018, 9:02 AM

Feb 8 2018

stefanp created D43086: [PowerPC] Infrastructure work. Implement getting the opcode for a spill in one place..
Feb 8 2018, 12:50 PM

Feb 5 2018

stefanp requested review of D39860: [PowerPC] Simplify a Swap if it feeds a Splat.
Feb 5 2018, 9:07 AM
stefanp updated the diff for D39860: [PowerPC] Simplify a Swap if it feeds a Splat.

Fixed the issue where we could have assigned a VSR register to an instruction that requires a VR register.

Feb 5 2018, 9:04 AM

Feb 1 2018

stefanp updated the diff for D40855: [PowerPC] LSR tunings for PowerPC.

Modified the tuning formula to take into consideration the idea that we shouldn’t be enlarging the class of equal items.
After looking into the LSR algorithm we discovered that if we enlarge the space of equal items we do not favor unchanged code as we had initially desired.
Therefore, this change does almost exactly what the default isLSRCostLess function does (from the base class) except that number of instructions has been added as the most important category to compare.

Feb 1 2018, 12:29 PM

Jan 31 2018

stefanp updated the diff for D42590: [PowerPC] Try to move the stack pointer update instruction later in the prologue and earlier in the epilogue (Version 2).

Fixed the issues mentioned in previous comments.

Jan 31 2018, 1:39 PM

Jan 26 2018

stefanp created D42590: [PowerPC] Try to move the stack pointer update instruction later in the prologue and earlier in the epilogue (Version 2).
Jan 26 2018, 11:04 AM

Jan 16 2018

stefanp added a comment to D40855: [PowerPC] LSR tunings for PowerPC.

@hfinkel
Hi Hal,
Yes that is correct. We are increasing the size of the class of equal elements.
The idea is that LSR will only replace one formula with another if it finds something with a lower cost. By returning false more often LSR will do fewer replacements and that is what we mean by more conservative. Maybe that's not really clear in the comment and I can update the comment to make it more clear.

Jan 16 2018, 2:09 PM

Jan 12 2018

stefanp commandeered D40855: [PowerPC] LSR tunings for PowerPC.

I'm going to start looking at this patch now.

Jan 12 2018, 10:51 AM
stefanp committed rL322369: Revert "[PowerPC] Manually schedule the prologue and epilogue".
Revert "[PowerPC] Manually schedule the prologue and epilogue"
Jan 12 2018, 5:14 AM

Jan 9 2018

stefanp committed rL322124: [PowerPC] Manually schedule the prologue and epilogue.
[PowerPC] Manually schedule the prologue and epilogue
Jan 9 2018, 1:59 PM

Jan 8 2018

stefanp committed rL322051: Revert "[PowerPC] Manually schedule the prologue and epilogue".
Revert "[PowerPC] Manually schedule the prologue and epilogue"
Jan 8 2018, 5:07 PM
stefanp committed rL322036: [PowerPC] Manually schedule the prologue and epilogue.
[PowerPC] Manually schedule the prologue and epilogue
Jan 8 2018, 2:24 PM
stefanp closed D41737: [PowerPC] Try to move the stack pointer update instruction later in the prologue and earlier in the epilogue.
Jan 8 2018, 2:24 PM

Jan 5 2018

stefanp added a comment to D41737: [PowerPC] Try to move the stack pointer update instruction later in the prologue and earlier in the epilogue.

Yes, we are hand scheduling the prologue and epilogue.
However, this is a little more than moving instructions around. Since we are moving the stack pointer update instruction, spills that use the stack pointer may have to be updated with new offsets.

Jan 5 2018, 1:28 PM
stefanp updated the diff for D41737: [PowerPC] Try to move the stack pointer update instruction later in the prologue and earlier in the epilogue.

Fixed a typo and a missing set of brackets.

Jan 5 2018, 9:20 AM
stefanp added inline comments to D41737: [PowerPC] Try to move the stack pointer update instruction later in the prologue and earlier in the epilogue.
Jan 5 2018, 6:28 AM

Jan 4 2018

stefanp created D41737: [PowerPC] Try to move the stack pointer update instruction later in the prologue and earlier in the epilogue.
Jan 4 2018, 12:10 PM

Dec 20 2017

stefanp committed rL321205: [PowerPC] Added an assert to make sure that the MBBI iterator is valid..
[PowerPC] Added an assert to make sure that the MBBI iterator is valid.
Dec 20 2017, 11:08 AM
stefanp closed D41358: [PowerPC] Fix minor bug to avoid a failure in createTailCallBranchInstr.
Dec 20 2017, 11:08 AM
stefanp updated the diff for D41358: [PowerPC] Fix minor bug to avoid a failure in createTailCallBranchInstr.

@hfinkel
Hi Hal,
Thank you for the approval. After doing a bit more investigation I realized that the condition should always be true. Before calling this function we always check that this is a return block. Therefore, I've change it slightly so that the guard is now an assert.

Dec 20 2017, 9:08 AM

Dec 18 2017

stefanp created D41358: [PowerPC] Fix minor bug to avoid a failure in createTailCallBranchInstr.
Dec 18 2017, 9:03 AM

Dec 6 2017

stefanp added a comment to D39805: [Power9] Set MicroOpBufferSize for Power 9.

Hi Everyone,
Thank you for your comments.

Dec 6 2017, 11:40 AM

Nov 22 2017

stefanp added a comment to D39805: [Power9] Set MicroOpBufferSize for Power 9.

Hi Eric,

Nov 22 2017, 11:50 AM

Nov 9 2017

stefanp created D39860: [PowerPC] Simplify a Swap if it feeds a Splat.
Nov 9 2017, 12:18 PM

Nov 8 2017

stefanp created D39805: [Power9] Set MicroOpBufferSize for Power 9.
Nov 8 2017, 8:12 AM

Oct 30 2017

stefanp committed rL316952: Revert "[PowerPC] Try to simplify a Swap if it feeds a Splat".
Revert "[PowerPC] Try to simplify a Swap if it feeds a Splat"
Oct 30 2017, 12:56 PM

Oct 24 2017

stefanp committed rL316478: [PowerPC] Try to simplify a Swap if it feeds a Splat.
[PowerPC] Try to simplify a Swap if it feeds a Splat
Oct 24 2017, 10:44 AM

Oct 23 2017

stefanp committed rL316371: Revert "[PowerPC] Try to simplify a Swap if it feeds a Splat".
Revert "[PowerPC] Try to simplify a Swap if it feeds a Splat"
Oct 23 2017, 1:24 PM
stefanp committed rL316366: [PowerPC] Try to simplify a Swap if it feeds a Splat.
[PowerPC] Try to simplify a Swap if it feeds a Splat
Oct 23 2017, 12:35 PM
stefanp closed D39009: [PowerPC] Simplify a Swap if it feeds a Splat by committing rL316366: [PowerPC] Try to simplify a Swap if it feeds a Splat.
Oct 23 2017, 12:35 PM
stefanp updated the diff for D39009: [PowerPC] Simplify a Swap if it feeds a Splat.

Sorry for the late reply Eric. I was on vacation Thursday and Friday last week.

Oct 23 2017, 6:29 AM