Roman Lebedev
@lebedevri.bsky.social
Somewhat interested in permanently migrating to DE/CZ.
Consider resending mails from '22-03/'24-02.
https://github.com/LebedevRI
https://matrix.to/#/@LebedevRI:matrix.org
https://fosstodon.org/@LebedevRI
https://t.me/LebedevRI
Consider resending mails from '22-03/'24-02.
https://github.com/LebedevRI
https://matrix.to/#/@LebedevRI:matrix.org
https://fosstodon.org/@LebedevRI
https://t.me/LebedevRI
RIP, Dad. I can only hope your were happy in your life. May your next reincarnation perhaps be a higher one, or at least with a better life, whenever/wherever/whatever it is. Perhaps we may talk again, in my lifetime, even if only in an 'Astral' realm.
February 6, 2025 at 7:34 PM
RIP, Dad. I can only hope your were happy in your life. May your next reincarnation perhaps be a higher one, or at least with a better life, whenever/wherever/whatever it is. Perhaps we may talk again, in my lifetime, even if only in an 'Astral' realm.
#PipeWire 25 bits is enough for everyone. `@`PipewireP https://gitlab.freedesktop.org/pipewire/pipewire/-/merge_requests/2040
February 6, 2025 at 7:34 PM
#PipeWire 25 bits is enough for everyone. `@`PipewireP https://gitlab.freedesktop.org/pipewire/pipewire/-/merge_requests/2040
Better late than never? Coming to a #NinjaBuildSystem near you, at last! Kudos to `@`esgamesorg! https://github.com/ninja-build/ninja/pull/1963 `@`phoronix
February 6, 2025 at 7:34 PM
Better late than never? Coming to a #NinjaBuildSystem near you, at last! Kudos to `@`esgamesorg! https://github.com/ninja-build/ninja/pull/1963 `@`phoronix
#EBU #R128 #EBUR128 Best audio player? https://github.com/strawberrymusicplayer/strawberry/pull/1216
February 6, 2025 at 7:34 PM
#EBU #R128 #EBUR128 Best audio player? https://github.com/strawberrymusicplayer/strawberry/pull/1216
`@`llvmorg `@`SylvestreLedru `@`alexreinking Fascinating times :) #LLVMWorldDomination https://x.com/DebianUpload/status/1462483485663367174
February 6, 2025 at 7:33 PM
`@`llvmorg `@`SylvestreLedru `@`alexreinking Fascinating times :) #LLVMWorldDomination https://x.com/DebianUpload/status/1462483485663367174
RT `@`DebianNew: New package: halide Version: 12.0.1-1 by Roman Lebedev ... http://deb.li/3tPdA
February 6, 2025 at 7:33 PM
RT `@`DebianNew: New package: halide Version: 12.0.1-1 by Roman Lebedev ... http://deb.li/3tPdA
`@`llvmorg `@`AMD `@`AMDRyzen #Zen3 #LLVM #Clang Interesting results, showing just how much proper per-CPU tuning matters. Free compute performance boost! You're welcome? https://x.com/phoronix/status/1425892252463910928
February 6, 2025 at 7:33 PM
`@`llvmorg `@`AMD `@`AMDRyzen #Zen3 #LLVM #Clang Interesting results, showing just how much proper per-CPU tuning matters. Free compute performance boost! You're welcome? https://x.com/phoronix/status/1425892252463910928
`@`llvmorg `@`llvmweekly `@`phoronix `@`AMD `@`AMDRyzen
6 months for `-march=znver3` is better than 6 years it took for `-march=bdver2` :) Though, it actually only took me ~42 days since i started needing said sched model. https://x.com/phoronix/status/1389175735101833217
6 months for `-march=znver3` is better than 6 years it took for `-march=bdver2` :) Though, it actually only took me ~42 days since i started needing said sched model. https://x.com/phoronix/status/1389175735101833217
February 6, 2025 at 7:33 PM
`@`llvmorg `@`llvmweekly `@`phoronix `@`AMD `@`AMDRyzen
6 months for `-march=znver3` is better than 6 years it took for `-march=bdver2` :) Though, it actually only took me ~42 days since i started needing said sched model. https://x.com/phoronix/status/1389175735101833217
6 months for `-march=znver3` is better than 6 years it took for `-march=bdver2` :) Though, it actually only took me ~42 days since i started needing said sched model. https://x.com/phoronix/status/1389175735101833217
`@`johnregehr That transform was a fun one to come up with!
Gaining a way to actually prove it (and that was made possible about the same time) was perhaps even more awesome, thanks `@`vpshufb & others!
Gaining a way to actually prove it (and that was made possible about the same time) was perhaps even more awesome, thanks `@`vpshufb & others!
February 6, 2025 at 7:33 PM
`@`johnregehr That transform was a fun one to come up with!
Gaining a way to actually prove it (and that was made possible about the same time) was perhaps even more awesome, thanks `@`vpshufb & others!
Gaining a way to actually prove it (and that was made possible about the same time) was perhaps even more awesome, thanks `@`vpshufb & others!
Achievement unlocked: branch predictors are fun https://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20200907/827578.html
February 6, 2025 at 7:33 PM
Achievement unlocked: branch predictors are fun https://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20200907/827578.html
`@`johnregehr Since there is already structure-aware LLVM IR fuzzing, the very obvious next step would be to integrate alive2 into instcombine fuzzer - having a way to tell if IR transformation is a miscompile or not is great, but it requires preexisting IR to function..
February 6, 2025 at 7:33 PM
`@`johnregehr Since there is already structure-aware LLVM IR fuzzing, the very obvious next step would be to integrate alive2 into instcombine fuzzer - having a way to tell if IR transformation is a miscompile or not is great, but it requires preexisting IR to function..
`@`johnregehr Circling back to this, increment/decrement support done too: https://github.com/llvm/llvm-project/commit/b98a0c7f6c9c7b38b6cd764e6f47e16e3d8c342c
i think mostly only bit field handling is missing..
(did not check volatile/atomic/enum status)
i think mostly only bit field handling is missing..
(did not check volatile/atomic/enum status)
February 6, 2025 at 7:33 PM
`@`johnregehr Circling back to this, increment/decrement support done too: https://github.com/llvm/llvm-project/commit/b98a0c7f6c9c7b38b6cd764e6f47e16e3d8c342c
i think mostly only bit field handling is missing..
(did not check volatile/atomic/enum status)
i think mostly only bit field handling is missing..
(did not check volatile/atomic/enum status)
More UB fun, with live miscompilations in-the-wild: applying non-zero (in C - any) offset to null pointer is undefined behavior in C++
https://github.com/llvm/llvm-project/commit/536b0ee40ab97f2878dc124a321cf9108ee3d233 `@`johnregehr `@`kayseesee `@`Philip_Reames #llvm #clang #ubsan
https://github.com/llvm/llvm-project/commit/536b0ee40ab97f2878dc124a321cf9108ee3d233 `@`johnregehr `@`kayseesee `@`Philip_Reames #llvm #clang #ubsan
February 6, 2025 at 7:33 PM
More UB fun, with live miscompilations in-the-wild: applying non-zero (in C - any) offset to null pointer is undefined behavior in C++
https://github.com/llvm/llvm-project/commit/536b0ee40ab97f2878dc124a321cf9108ee3d233 `@`johnregehr `@`kayseesee `@`Philip_Reames #llvm #clang #ubsan
https://github.com/llvm/llvm-project/commit/536b0ee40ab97f2878dc124a321cf9108ee3d233 `@`johnregehr `@`kayseesee `@`Philip_Reames #llvm #clang #ubsan
Who doesn't love benchmarking? `@`runtime https://x.com/GoogleOSS/status/1121825826939424768
February 6, 2025 at 7:33 PM
Who doesn't love benchmarking? `@`runtime https://x.com/GoogleOSS/status/1121825826939424768
Thanks for the opportunity, it was exciting! https://x.com/johnregehr/status/1083434978845777920
February 6, 2025 at 7:33 PM
Thanks for the opportunity, it was exciting! https://x.com/johnregehr/status/1083434978845777920
`@`johnregehr Nice! Sorry about that, fixed in https://reviews.llvm.org/rL348181
Really looking forward souper-driven next-gen instcombine
without any room for such issues :)
Really looking forward souper-driven next-gen instcombine
without any room for such issues :)
February 6, 2025 at 7:33 PM
`@`johnregehr Nice! Sorry about that, fixed in https://reviews.llvm.org/rL348181
Really looking forward souper-driven next-gen instcombine
without any room for such issues :)
Really looking forward souper-driven next-gen instcombine
without any room for such issues :)
`@`ON1photo `@`ON1_Dan `@`darixzen `@`cryptomilk `@`pixlsus `@`fsf `@`sh0dan 1. it's always "darktable" :)
2. as it can be seen from the repo history, the license has always been the exactly the same, so the question still stands unanswered.
> Many thanks to `@`sh0dan for his hard work!
+1
2. as it can be seen from the repo history, the license has always been the exactly the same, so the question still stands unanswered.
> Many thanks to `@`sh0dan for his hard work!
+1
February 6, 2025 at 7:33 PM
`@`ON1photo `@`ON1_Dan `@`darixzen `@`cryptomilk `@`pixlsus `@`fsf `@`sh0dan 1. it's always "darktable" :)
2. as it can be seen from the repo history, the license has always been the exactly the same, so the question still stands unanswered.
> Many thanks to `@`sh0dan for his hard work!
+1
2. as it can be seen from the repo history, the license has always been the exactly the same, so the question still stands unanswered.
> Many thanks to `@`sh0dan for his hard work!
+1
`@`johnregehr https://reviews.llvm.org/rL347258
February 6, 2025 at 7:33 PM
`@`johnregehr https://reviews.llvm.org/rL347258
`@`johnregehr `@`gregbedwell
https://llvm.org/devmtg/2018-04/slides/Bedwell-Measuring_the_User_Debugging_Experience.pdf
https://lists.llvm.org/pipermail/llvm-dev/2018-August/125780.html
https://llvm.org/devmtg/2018-04/slides/Bedwell-Measuring_the_User_Debugging_Experience.pdf
https://lists.llvm.org/pipermail/llvm-dev/2018-August/125780.html
February 6, 2025 at 7:33 PM
`@`johnregehr `@`gregbedwell
https://llvm.org/devmtg/2018-04/slides/Bedwell-Measuring_the_User_Debugging_Experience.pdf
https://lists.llvm.org/pipermail/llvm-dev/2018-August/125780.html
https://llvm.org/devmtg/2018-04/slides/Bedwell-Measuring_the_User_Debugging_Experience.pdf
https://lists.llvm.org/pipermail/llvm-dev/2018-August/125780.html
`@`johnregehr https://reviews.llvm.org/D53949 *should* do it, although needs tests.
February 6, 2025 at 7:33 PM
`@`johnregehr https://reviews.llvm.org/D53949 *should* do it, although needs tests.
`@`johnregehr Nice. No, that is not intentional, it is a bug. As it can be seen from the AST https://godbolt.org/z/0B8YZc the implicit downcast isn't actually present in the AST, at least as a separate node, but is implicitly done by CompoundAssignOperator i would say. I'll take a look.
February 6, 2025 at 7:33 PM
`@`johnregehr Nice. No, that is not intentional, it is a bug. As it can be seen from the AST https://godbolt.org/z/0B8YZc the implicit downcast isn't actually present in the AST, at least as a separate node, but is implicitly done by CompoundAssignOperator i would say. I'll take a look.
`@`johnregehr `@`wdtz Note that it is the second half - sign change catching.
The first part was https://reviews.llvm.org/D48958 - data loss catching.
It's a complete write-from-scratch. I never saw any of the previous work.
The *only* remaining piece after that is bitfield handling.
The first part was https://reviews.llvm.org/D48958 - data loss catching.
It's a complete write-from-scratch. I never saw any of the previous work.
The *only* remaining piece after that is bitfield handling.
February 6, 2025 at 7:33 PM
`@`johnregehr `@`wdtz Note that it is the second half - sign change catching.
The first part was https://reviews.llvm.org/D48958 - data loss catching.
It's a complete write-from-scratch. I never saw any of the previous work.
The *only* remaining piece after that is bitfield handling.
The first part was https://reviews.llvm.org/D48958 - data loss catching.
It's a complete write-from-scratch. I never saw any of the previous work.
The *only* remaining piece after that is bitfield handling.
`@`DuaneOBrien `@`freenodestaff `@`DuaneOBrien well, that is an interesting way to react :)
February 6, 2025 at 7:33 PM
`@`DuaneOBrien `@`freenodestaff `@`DuaneOBrien well, that is an interesting way to react :)
`@`autoexec_bin `@`patdavid `@`pixlsus RPU already has such a sample https://raw.pixls.us/getfile.php/980/raw/B13A0729.CR2
February 6, 2025 at 7:33 PM
`@`autoexec_bin `@`patdavid `@`pixlsus RPU already has such a sample https://raw.pixls.us/getfile.php/980/raw/B13A0729.CR2
`@`phoronix `@`openclapi `@`MichaelLarabel `@`darktable_org it's darktable. NOT DarkTable/Darktable/etc
February 6, 2025 at 7:33 PM
`@`phoronix `@`openclapi `@`MichaelLarabel `@`darktable_org it's darktable. NOT DarkTable/Darktable/etc