[PATCH] D149776: Re-land "[AMDGPU] Define data layout entries for buffers""

krzysz00 added a comment. From a quick skim of the backtrace, I think this comes down to the fact that there isn't a defined way to stick a `p7` into registers (or, equivalently, there's no `MVT::i160`) and that getting said pointers to that point was supposed to never happen (in that they were

[PATCH] D149776: Re-land "[AMDGPU] Define data layout entries for buffers""

This revision was not accepted when it landed; it landed in state "Needs Review". This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rGf0415f2a456d: Re-land "[AMDGPU] Define data layout entries for buf

[PATCH] D149776: Re-land "[AMDGPU] Define data layout entries for buffers""

krzysz00 created this revision. Herald added subscribers: kosarev, jeroen.dobbelaere, foad, wenlei, okura, kuter, kerbowa, arphaman, zzheng, hiraditya, tpr, dstuttard, yaxunl, jvesely, kzhuravl, arsenm, MatzeB. Herald added a project: All. krzysz00 requested review of this revision. Herald added

[PATCH] D149758: Revert "[AMDGPU] Define data layout entries for buffers"

This revision was not accepted when it landed; it landed in state "Needs Review". This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rG3f2fbe92d0f4: Revert "[AMDGPU] Define data layout entries for buff

[PATCH] D149758: Revert "[AMDGPU] Define data layout entries for buffers"

krzysz00 created this revision. Herald added subscribers: kosarev, jeroen.dobbelaere, foad, wenlei, okura, kuter, kerbowa, arphaman, zzheng, hiraditya, tpr, dstuttard, yaxunl, jvesely, kzhuravl, arsenm, MatzeB. Herald added a project: All. krzysz00 requested review of this revision. Herald added

[PATCH] D145441: [AMDGPU] Define data layout entries for buffers

This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rGf9c1ede2543b: [AMDGPU] Define data layout entries for buffers (authored by krzysz00). Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTI

[PATCH] D145441: [AMDGPU] Define data layout entries for buffers

krzysz00 updated this revision to Diff 518843. krzysz00 added a comment. Rebase to handle new legalization tests. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D145441/new/ https://reviews.llvm.org/D145441 Files: clang/lib/Basic/Targets/AMDGPU.cp

[PATCH] D145441: [AMDGPU] Define data layout entries for buffers

krzysz00 updated this revision to Diff 516953. krzysz00 edited the summary of this revision. krzysz00 added a comment. Make address space 8 non-integral as well because that's a good idea and we'll have an address space cast intrinsic. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST A

[PATCH] D145441: [AMDGPU] Define data layout entries for buffers

krzysz00 updated this revision to Diff 514232. krzysz00 added a comment. Fix new merge conflicts Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D145441/new/ https://reviews.llvm.org/D145441 Files: clang/lib/Basic/Targets/AMDGPU.cpp clang/test/Co

[PATCH] D145441: [AMDGPU] Define data layout entries for buffers

krzysz00 updated this revision to Diff 512488. krzysz00 added a comment. Rebase, since we have addrspace 128 Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D145441/new/ https://reviews.llvm.org/D145441 Files: clang/lib/Basic/Targets/AMDGPU.cpp c

[PATCH] D145441: [AMDGPU] Define data layout entries for buffers

krzysz00 updated this revision to Diff 510848. krzysz00 added a comment. Rebase data layout changes Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D145441/new/ https://reviews.llvm.org/D145441 Files: clang/lib/Basic/Targets/AMDGPU.cpp clang/test

[PATCH] D145441: [AMDGPU] Define data layout entries for buffers

krzysz00 updated this revision to Diff 510056. krzysz00 edited the summary of this revision. krzysz00 added a comment. Per discussion on the s.buffer.load revision, don't make those changes, and so revert this patch back to what it used to be. Repository: rG LLVM Github Monorepo CHANGES SINC

[PATCH] D145441: [AMDGPU] Define data layout entries for buffers

krzysz00 updated this revision to Diff 509761. krzysz00 edited the summary of this revision. krzysz00 added a comment. Rebase for splitting s.buffer.load effect change into earlier patch Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D145441/new/ htt

[PATCH] D145441: [AMDGPU] Define data layout entries for buffers

krzysz00 updated this revision to Diff 507870. krzysz00 added a comment. ... actually fix the typo Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D145441/new/ https://reviews.llvm.org/D145441 Files: clang/lib/Basic/Targets/AMDGPU.cpp clang/test/

[PATCH] D145441: [AMDGPU] Define data layout entries for buffers

krzysz00 updated this revision to Diff 507869. krzysz00 added a comment. Fix typo, rebase while making next patch Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D145441/new/ https://reviews.llvm.org/D145441 Files: clang/lib/Basic/Targets/AMDGPU.cp

[PATCH] D145441: [AMDGPU] Define data layout entries for buffers

krzysz00 added a comment. @foad I was trying to avoid sending in one mega-patch that has the entire prototype. As to your comments: - Why 256-bit alignment? It's the next power of 2, and using an alignment that isn't a power of 2 causes an assertion failure - Re not being allowed to use addres

[PATCH] D145441: [AMDGPU] Define data layout entries for buffers

krzysz00 created this revision. krzysz00 added reviewers: nhaehnle, arsenm, b-sumner, piotr, sstefan1, jdoerfert. Herald added subscribers: kosarev, jeroen.dobbelaere, foad, wenlei, okura, kuter, kerbowa, arphaman, zzheng, hiraditya, arichardson, tpr, dstuttard, yaxunl, jvesely, kzhuravl, MatzeB

[PATCH] D143522: [AMDGPU] Set a data layout entry for buffer descriptors (addrspace 7)

krzysz00 added a reviewer: piotr. krzysz00 added a comment. Discussion related to this is also happening over at https://reviews.llvm.org/D143945 . Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D143522/new/ https://reviews.llvm.org/D143522 ___

[PATCH] D143522: [AMDGPU] Set a data layout entry for buffer descriptors (addrspace 7)

krzysz00 updated this revision to Diff 496515. krzysz00 added a comment. Herald added a project: clang. Herald added a subscriber: cfe-commits. Apparently clang has its own copy of this data layout. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D1435

[PATCH] D143684: Add missing cases to clang switch after D141863

This revision was not accepted when it landed; it landed in state "Needs Review". This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rG5d8da5a208e6: Add missing cases to clang switch after D141863 (aut

[PATCH] D143684: Add missing cases to clang switch after D141863

krzysz00 created this revision. Herald added a project: All. krzysz00 requested review of this revision. Herald added a project: clang. Herald added a subscriber: cfe-commits. Turns out there's a switch on APFloat semantics in clang I wasn't aware of, fix the build error here. Repository: rG L

[PATCH] D141432: Add two additional float8 types to MLIR and APFloat.

krzysz00 added a comment. Herald added a reviewer: ftynse. I think we have dueling patches that we should reconcile. I've put together https://reviews.llvm.org/D141863 to do the same thing. (it's just now posted because we were waiting to talk to y'all over at GraphCore to finalize naming, etc.

[PATCH] D129308: [mlir] Remove VectorToROCDL

This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rGd6ef3d20b4e3: [mlir] Remove VectorToROCDL (authored by krzysz00). Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews

[PATCH] D129308: [mlir] Remove VectorToROCDL

krzysz00 created this revision. Herald added subscribers: bzcheeseman, kosarev, sdasgup3, wenzhicui, wrengr, dcaballe, cota, teijeong, rdzhabarov, tatianashp, msifontes, jurahul, Kayjukh, grosul1, Joonsoo, liufengdb, aartbik, mgester, arpith-jacob, csigg, antiagainst, shauheen, rriddle, mehdi_am

[PATCH] D119257: Revert "Re-land [LLD] Remove global state in lldCommon"

krzysz00 abandoned this revision. krzysz00 added a comment. Worked around elsewhere, more permanent fix pending, abandoning Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D119257/new/ https://reviews.llvm.org/D119257

[PATCH] D119257: Revert "Re-land [LLD] Remove global state in lldCommon"

krzysz00 added a comment. Having checked with other folks more familiar with the context of our project, it looks like we might actually be fine using lld as a subprocess, since we'll always be sitting on top of a ROCm installation, which includes lld, and since "take this object file and wrap

[PATCH] D119257: Revert "Re-land [LLD] Remove global state in lldCommon"

krzysz00 added a comment. (Since it looks like the discussion moved here) @MaskRay We can't spawn lld as a subprocess due to customer requirements around having the kernel generator/compiler delivered as a static library in some cases. (And there's LLVM version compatibility concerns, but those

[PATCH] D119257: Revert "Re-land [LLD] Remove global state in lldCommon"

krzysz00 added a comment. @aganea Can we expect fixes for both the issues with this revision (the HIP-on-Windows problems noted last night and the MLIR breakage) shortly or should we revert while y'all work out fixes? Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://r

[PATCH] D119257: Revert "Re-land [LLD] Remove global state in lldCommon"

krzysz00 created this revision. krzysz00 added reviewers: aganea, MaskRay. Herald added subscribers: sdasgup3, wenzhicui, wrengr, Chia-hungDuan, ormris, dcaballe, cota, teijeong, rdzhabarov, tatianashp, msifontes, jurahul, Kayjukh, grosul1, Joonsoo, liufengdb, aartbik, mgester, arpith-jacob, csig