-
Notifications
You must be signed in to change notification settings - Fork 397
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Build fails #14352
Comments
I got the same error.
|
It fails for me with a different error, but still fails. I'll take a look whether it's something recent or it's broken for some time now. |
The build of this particular file is broken since end of May. There are two more similar cases, when GPU files fail to build in CPU build. If you try to build useful targets (rather than everything, I'll ask around what is the intended behaviour there, but likely it's one of the following:
|
Thanks @mooskagh. What is the command to build for cpu then? How do I check what's tested in CI to know what should be working? |
I tried
I see only certain parts of gloo are excluded for macos build (this not being one of them), so I assume this should work? Is there anything else I need to set up in my env? |
Can you please try whether a49d8aa fixes the issue with compiling command_buffer_cmd_emitter.cc ? |
@akuegel, what is the build command I should use? Just running
|
But that is a different error from before, so at least the other error is fixed now. Will look into the new error next week, unless someone else fixes it before then. |
Sounds good, thanks! |
cc2ec3f hopefully fixes this error related to not finding the gpu_sort_rewriter.h header |
Thanks @akuegel! Those initial errors seem to have been resolved. I'm running into this now:
Any thoughts? |
@tomnatan30 as this is a compile error in shardy, maybe you can take a look? |
Hey, I am also following similar instructions to build xla from source for the CPU backend and ran into the following error:
Am I missing anything? Thanks. |
@shardonnay-dev ***@***.***>
Thanks Adrian!
We will take a look
…On Thu, Jul 11, 2024 at 5:42 PM Hugh ***@***.***> wrote:
Hey, I am also following similar instructions
<https://github.com/openxla/xla/blob/main/docs/build_from_source.md> to
build xla from source for the CPU backend and ran into the following error:
In file included from ./xla/service/memory_space_assignment/memory_space_assignment.h:188:
./xla/service/memory_space_assignment/cost_analysis.h:118:3: warning: explicitly defaulted default constructor is implicitly deleted [-Wdefaulted-function-deleted]
118 | HloCostAnalysisCosts() = default;
| ^
./xla/service/memory_space_assignment/cost_analysis.h:120:26: note: default constructor of 'HloCostAnalysisCosts' is implicitly deleted because field 'hlo_cost_analysis_' of reference type 'const HloCostAnalysis &' would not be initialized
120 | const HloCostAnalysis& hlo_cost_analysis_;
| ^
./xla/service/memory_space_assignment/cost_analysis.h:118:28: note: replace 'default' with 'delete'
118 | HloCostAnalysisCosts() = default;
| ^~~~~~~
| delete
xla/service/gpu/fusions/triton.cc:41:10: fatal error: 'xla/service/gpu/ir_emitter_triton.h' file not found
41 | #include "xla/service/gpu/ir_emitter_triton.h"
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Am I missing anything? Thanks.
—
Reply to this email directly, view it on GitHub
<#14352 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/A7DIFH6LSKXKALXE36HL7RDZL2YWFAVCNFSM6AAAAABKGSDSNGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMRTGQYDQMRQGI>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
@ryan Lefever ***@***.***>
On Thu, Jul 11, 2024 at 9:55 AM 'Tom Natan' via shardonnay-dev <
***@***.***> wrote:
… @shardonnay-dev ***@***.***>
Thanks Adrian!
We will take a look
On Thu, Jul 11, 2024 at 5:42 PM Hugh ***@***.***> wrote:
> Hey, I am also following similar instructions
> <https://github.com/openxla/xla/blob/main/docs/build_from_source.md> to
> build xla from source for the CPU backend and ran into the following error:
>
> In file included from ./xla/service/memory_space_assignment/memory_space_assignment.h:188:
> ./xla/service/memory_space_assignment/cost_analysis.h:118:3: warning: explicitly defaulted default constructor is implicitly deleted [-Wdefaulted-function-deleted]
> 118 | HloCostAnalysisCosts() = default;
> | ^
> ./xla/service/memory_space_assignment/cost_analysis.h:120:26: note: default constructor of 'HloCostAnalysisCosts' is implicitly deleted because field 'hlo_cost_analysis_' of reference type 'const HloCostAnalysis &' would not be initialized
> 120 | const HloCostAnalysis& hlo_cost_analysis_;
> | ^
> ./xla/service/memory_space_assignment/cost_analysis.h:118:28: note: replace 'default' with 'delete'
> 118 | HloCostAnalysisCosts() = default;
> | ^~~~~~~
> | delete
> xla/service/gpu/fusions/triton.cc:41:10: fatal error: 'xla/service/gpu/ir_emitter_triton.h' file not found
> 41 | #include "xla/service/gpu/ir_emitter_triton.h"
> | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
> Am I missing anything? Thanks.
>
> —
> Reply to this email directly, view it on GitHub
> <#14352 (comment)>,
> or unsubscribe
> <https://github.com/notifications/unsubscribe-auth/A7DIFH6LSKXKALXE36HL7RDZL2YWFAVCNFSM6AAAAABKGSDSNGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMRTGQYDQMRQGI>
> .
> You are receiving this because you were mentioned.Message ID:
> ***@***.***>
>
--
You received this message because you are subscribed to the Google Groups
"shardonnay-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to ***@***.***
To view this discussion on the web visit
https://groups.google.com/a/google.com/d/msgid/shardonnay-dev/CADw3ZZ4PfibnqwKSyyF2RMrkkhXYkyTccxG59O-50v-BN_7ADQ%40mail.gmail.com
<https://groups.google.com/a/google.com/d/msgid/shardonnay-dev/CADw3ZZ4PfibnqwKSyyF2RMrkkhXYkyTccxG59O-50v-BN_7ADQ%40mail.gmail.com?utm_medium=email&utm_source=footer>
.
For more options, visit https://groups.google.com/a/google.com/d/optout.
|
This might be a c++ version issue, it looks like structured binding that we
have in the code. This wasn't a problem for us when building the Shardy
repo using bazel, could this be related to the specific command you are
using to build XLA?
…On Thu, Jul 11, 2024 at 5:57 PM Andy Davis ***@***.***> wrote:
@ryan Lefever ***@***.***>
On Thu, Jul 11, 2024 at 9:55 AM 'Tom Natan' via shardonnay-dev <
***@***.***> wrote:
> @shardonnay-dev ***@***.***>
>
> Thanks Adrian!
>
> We will take a look
>
> On Thu, Jul 11, 2024 at 5:42 PM Hugh ***@***.***> wrote:
>
>> Hey, I am also following similar instructions
>> <https://github.com/openxla/xla/blob/main/docs/build_from_source.md> to
>> build xla from source for the CPU backend and ran into the following error:
>>
>> In file included from ./xla/service/memory_space_assignment/memory_space_assignment.h:188:
>> ./xla/service/memory_space_assignment/cost_analysis.h:118:3: warning: explicitly defaulted default constructor is implicitly deleted [-Wdefaulted-function-deleted]
>> 118 | HloCostAnalysisCosts() = default;
>> | ^
>> ./xla/service/memory_space_assignment/cost_analysis.h:120:26: note: default constructor of 'HloCostAnalysisCosts' is implicitly deleted because field 'hlo_cost_analysis_' of reference type 'const HloCostAnalysis &' would not be initialized
>> 120 | const HloCostAnalysis& hlo_cost_analysis_;
>> | ^
>> ./xla/service/memory_space_assignment/cost_analysis.h:118:28: note: replace 'default' with 'delete'
>> 118 | HloCostAnalysisCosts() = default;
>> | ^~~~~~~
>> | delete
>> xla/service/gpu/fusions/triton.cc:41:10: fatal error: 'xla/service/gpu/ir_emitter_triton.h' file not found
>> 41 | #include "xla/service/gpu/ir_emitter_triton.h"
>> | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>
>> Am I missing anything? Thanks.
>>
>> —
>> Reply to this email directly, view it on GitHub
>> <#14352 (comment)>,
>> or unsubscribe
>> <https://github.com/notifications/unsubscribe-auth/A7DIFH6LSKXKALXE36HL7RDZL2YWFAVCNFSM6AAAAABKGSDSNGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMRTGQYDQMRQGI>
>> .
>> You are receiving this because you were mentioned.Message ID:
>> ***@***.***>
>>
> --
> You received this message because you are subscribed to the Google Groups
> "shardonnay-dev" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to ***@***.***
> To view this discussion on the web visit
> https://groups.google.com/a/google.com/d/msgid/shardonnay-dev/CADw3ZZ4PfibnqwKSyyF2RMrkkhXYkyTccxG59O-50v-BN_7ADQ%40mail.gmail.com
> <https://groups.google.com/a/google.com/d/msgid/shardonnay-dev/CADw3ZZ4PfibnqwKSyyF2RMrkkhXYkyTccxG59O-50v-BN_7ADQ%40mail.gmail.com?utm_medium=email&utm_source=footer>
> .
> For more options, visit https://groups.google.com/a/google.com/d/optout.
>
--
You received this message because you are subscribed to the Google Groups
"shardonnay-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to ***@***.***
To view this discussion on the web visit
https://groups.google.com/a/google.com/d/msgid/shardonnay-dev/CAFfp4mP7ed0fX3EtKWMcXFpfJc0usnseNt2CheuQRf0rz0EP2A%40mail.gmail.com
<https://groups.google.com/a/google.com/d/msgid/shardonnay-dev/CAFfp4mP7ed0fX3EtKWMcXFpfJc0usnseNt2CheuQRf0rz0EP2A%40mail.gmail.com?utm_medium=email&utm_source=footer>
.
For more options, visit https://groups.google.com/a/google.com/d/optout.
|
I was just following the instruction on the build page:
|
I cloned https://github.com/openxla/shardy and built via `bazel build
--test_output=all --spawn_strategy=sandboxed //shardy/...` and it worked,
not sure about `./configure.py --backend=CPU` though. Can you try without
and see if it builds or if you get different errors?
…On Thu, Jul 11, 2024 at 6:05 PM Aleks Knezevic ***@***.***> wrote:
I was just following the instruction on the build page:
./configure.py --backend=CPU
bazel build --test_output=all --spawn_strategy=sandboxed //xla/...
—
Reply to this email directly, view it on GitHub
<#14352 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/A7DIFHZKJK3MEBFQIZ6IHHLZL23MBAVCNFSM6AAAAABKGSDSNGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMRTGQ2DQMJXGU>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
@tomnatan30 I think this might be a gcc specific compile error. We have seen that in similar cases in the XLA code base. The fix should be to explicitly capture valueIndex here: https://github.com/openxla/shardy/blob/main/shardy/dialect/sdy/ir/verifiers.cc#L706 |
openxla/shardy#8 should hopefully fix this error |
The fix has been merged in openxla/shardy@2da0ff5, @AleksKnezevic can you please give it another try? Edit: nevermind, I think I need to also update the pin of shardy in the xla repo. |
Thanks Adrian!
…On Fri, Jul 12, 2024 at 8:11 AM Adrian Kuegel ***@***.***> wrote:
The fix has been merged in ***@***.***
<openxla/shardy@2da0ff5>,
@AleksKnezevic <https://github.com/AleksKnezevic> can you please give it
another try?
—
Reply to this email directly, view it on GitHub
<#14352 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/A7DIFH2GSIGFBK6EI6GFBODZL56STAVCNFSM6AAAAABKGSDSNGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMRUHE3DMMRRGI>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
Thanks @akuegel, lmk once it's in, I tried updating it to test but I'm getting a but an
error. |
The Patch file is obsolete If you update to the latest revision. So you can just remove everything in shardy.patch. Unfortunately the change to update the revision requires approval from someone in the US. |
Awesome, building now. btw, what tool do you use to be able to see/browse third_party at the right commit and create these patches? I'm not too familiar with bazel, and used to git submodules |
Now running into:
|
I think the commit hash in shardy/workspace.bzl should be updated to
latest, we will update this soon and it should resolve the issue.
We plan on doing this automatically in the future.
…On Fri, Jul 12, 2024 at 3:58 PM Aleks Knezevic ***@***.***> wrote:
Now running into:
ERROR: /localdev/aknezevic/.cache/bazel/_bazel_aknezevic/f015ee4000db43b08516019840665d5e/external/shardy/shardy/dialect/sdy/ir/BUILD:113:11: Compiling shardy/dialect/sdy/ir/verifiers.cc failed: (Exit 1): clang failed: error executing command (from target @shardy//shardy/dialect/sdy/ir:dialect) /usr/lib/llvm-14/bin/clang -U_FORTIFY_SOURCE -fstack-protector -Wall -Wthread-safety -Wself-assign -Wunused-but-set-parameter -Wno-free-nonheap-object -fcolor-diagnostics -fno-omit-frame-pointer -g0 ... (remaining 121 arguments skipped)
external/shardy/shardy/dialect/sdy/ir/verifiers.cc:706:33: error: 'valueIndex' in capture list does not name a variable
auto emitManualAxesError = [valueIndex, &valueKindStr, &op](
^
external/shardy/shardy/dialect/sdy/ir/verifiers.cc:709:42: error: reference to local binding 'valueIndex' declared in enclosing function 'mlir::sdy::(anonymous namespace)::verifyManualComputationValue'
<< " sharding at index " << valueIndex
^
external/shardy/shardy/dialect/sdy/ir/verifiers.cc:665:14: note: 'valueIndex' declared here
for (auto [valueIndex, valueEntry] : llvm::enumerate(llvm::zip_equal(
^
external/shardy/shardy/dialect/sdy/ir/verifiers.cc:709:42: error: reference to local binding 'valueIndex' declared in enclosing function 'mlir::sdy::(anonymous namespace)::verifyManualComputationValue<mlir::ValueTypeRange<mlir::OperandRange>, mlir::ValueTypeRange<llvm::MutableArrayRef<mlir::BlockArgument>>>'
<< " sharding at index " << valueIndex
^
external/shardy/shardy/dialect/sdy/ir/verifiers.cc:789:14: note: in instantiation of function template specialization 'mlir::sdy::(anonymous namespace)::verifyManualComputationValue<mlir::ValueTypeRange<mlir::OperandRange>, mlir::ValueTypeRange<llvm::MutableArrayRef<mlir::BlockArgument>>>' requested here
if (failed(verifyManualComputationValue(
^
external/shardy/shardy/dialect/sdy/ir/verifiers.cc:665:14: note: 'valueIndex' declared here
for (auto [valueIndex, valueEntry] : llvm::enumerate(llvm::zip_equal(
^
external/shardy/shardy/dialect/sdy/ir/verifiers.cc:709:42: error: reference to local binding 'valueIndex' declared in enclosing function 'mlir::sdy::(anonymous namespace)::verifyManualComputationValue<mlir::ValueTypeRange<mlir::ResultRange>, mlir::TypeRange>'
<< " sharding at index " << valueIndex
^
external/shardy/shardy/dialect/sdy/ir/verifiers.cc:792:14: note: in instantiation of function template specialization 'mlir::sdy::(anonymous namespace)::verifyManualComputationValue<mlir::ValueTypeRange<mlir::ResultRange>, mlir::TypeRange>' requested here
failed(verifyManualComputationValue(
^
external/shardy/shardy/dialect/sdy/ir/verifiers.cc:665:14: note: 'valueIndex' declared here
for (auto [valueIndex, valueEntry] : llvm::enumerate(llvm::zip_equal(
^
—
Reply to this email directly, view it on GitHub
<#14352 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/A7DIFH2WJXJYZ6IKFPQ3TJLZL7VKJAVCNFSM6AAAAABKGSDSNGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMRVG43TIOJSGY>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
So this was with the latest commit in shardy:
|
Why is SHARDY_SHA256 empty?
Here is what I have:
SHARDY_COMMIT = "2fe34232a42752d8ae972948db9fd48a6f33f412"
SHARDY_SHA256 =
"7358eea0582347c3c8c51041535ac97536214e719a88e1655c724648d7ab70ea"
And then the patch file should be empty
…On Fri, Jul 12, 2024 at 5:34 PM Aleks Knezevic ***@***.***> wrote:
So this was with the latest commit in shardy:
/localdev/aknezevic/xla > git diff third_party/shardy/workspace.bzl
diff --git a/third_party/shardy/workspace.bzl b/third_party/shardy/workspace.bzl
index ca3f6421dc..29bdd6f908 100644
--- a/third_party/shardy/workspace.bzl
+++ b/third_party/shardy/workspace.bzl
@@ -3,13 +3,12 @@
load("//third_party:repo.bzl", "tf_http_archive", "tf_mirror_urls")
def repo():
- SHARDY_COMMIT = "7afabee9bf7addaef719244fe0a605463738384d"
- SHARDY_SHA256 = "7271375db347541dfd544fca1390b9aa700013849fba22c48b46bcf04a7f281a"
+ SHARDY_COMMIT = "2fe34232a42752d8ae972948db9fd48a6f33f412"
+ SHARDY_SHA256 = ""
tf_http_archive(
name = "shardy",
sha256 = SHARDY_SHA256,
strip_prefix = "shardy-{commit}".format(commit = SHARDY_COMMIT),
urls = tf_mirror_urls("https://github.com/openxla/shardy/archive/{commit}.zip".format(commit = SHARDY_COMMIT)),
- patch_file = ["//third_party/shardy:shardy.patch"],
)
—
Reply to this email directly, view it on GitHub
<#14352 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/A7DIFH6NLZNQZOCEULNJA4TZMAASFAVCNFSM6AAAAABKGSDSNGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMRVHEZTKMRXGU>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
And empty SHA256 just ignores the check, I believe, and I removed the patch_file line. The correct commit is applied, if you look at the errors above, they're different. |
Oh I think I know what the fix might be
…On Fri, Jul 12, 2024 at 5:38 PM Aleks Knezevic ***@***.***> wrote:
And empty SHA256 just ignores the check, I believe, and I removed the
patch_file line. The correct commit is applied, if you look at the errors
above, they're different.
—
Reply to this email directly, view it on GitHub
<#14352 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/A7DIFH63XEB3RSOPUF7HF43ZMABBPAVCNFSM6AAAAABKGSDSNGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMRVHE2DAOBVHA>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
Sending a fix now.
…On Fri, Jul 12, 2024 at 5:39 PM Tom Natan ***@***.***> wrote:
Oh I think I know what the fix might be
On Fri, Jul 12, 2024 at 5:38 PM Aleks Knezevic ***@***.***>
wrote:
> And empty SHA256 just ignores the check, I believe, and I removed the
> patch_file line. The correct commit is applied, if you look at the errors
> above, they're different.
>
> —
> Reply to this email directly, view it on GitHub
> <#14352 (comment)>,
> or unsubscribe
> <https://github.com/notifications/unsubscribe-auth/A7DIFH63XEB3RSOPUF7HF43ZMABBPAVCNFSM6AAAAABKGSDSNGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMRVHE2DAOBVHA>
> .
> You are receiving this because you were mentioned.Message ID:
> ***@***.***>
>
|
Can you please try with the newest commit?
…On Fri, Jul 12, 2024 at 5:49 PM Tom Natan ***@***.***> wrote:
Sending a fix now.
On Fri, Jul 12, 2024 at 5:39 PM Tom Natan ***@***.***> wrote:
> Oh I think I know what the fix might be
>
> On Fri, Jul 12, 2024 at 5:38 PM Aleks Knezevic ***@***.***>
> wrote:
>
>> And empty SHA256 just ignores the check, I believe, and I removed the
>> patch_file line. The correct commit is applied, if you look at the errors
>> above, they're different.
>>
>> —
>> Reply to this email directly, view it on GitHub
>> <#14352 (comment)>,
>> or unsubscribe
>> <https://github.com/notifications/unsubscribe-auth/A7DIFH63XEB3RSOPUF7HF43ZMABBPAVCNFSM6AAAAABKGSDSNGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMRVHE2DAOBVHA>
>> .
>> You are receiving this because you were mentioned.Message ID:
>> ***@***.***>
>>
>
|
@AleksKnezevic I can compile xla successfully now with @akuegel Thanks for the quick fix in the gpu directory! |
@tomnatan30 on the latest commit I hit:
|
Sorry about that, can you try bumping again to the latest commit -
openxla/shardy#15
…On Fri, Jul 12, 2024 at 9:47 PM Aleks Knezevic ***@***.***> wrote:
@tomnatan30 <https://github.com/tomnatan30> on the latest commit I hit:
ERROR: /localdev/aknezevic/.cache/bazel/_bazel_aknezevic/f015ee4000db43b08516019840665d5e/external/shardy/shardy/dialect/sdy/transforms/propagation/BUILD:93:11: Compiling shardy/dialect/sdy/transforms/propagation/op_sharding_rule_registry.cc failed: (Exit 1): clang failed: error executing command (from target @shardy//shardy/dialect/sdy/transforms/propagation:op_sharding_rule_registry) /usr/lib/llvm-14/bin/clang -U_FORTIFY_SOURCE -fstack-protector -Wall -Wthread-safety -Wself-assign -Wunused-but-set-parameter -Wno-free-nonheap-object -fcolor-diagnostics -fno-omit-frame-pointer -g0 ... (remaining 121 arguments skipped)
Use --sandbox_debug to see verbose messages from the sandbox and retain the sandbox build root for debugging
external/shardy/shardy/dialect/sdy/transforms/propagation/op_sharding_rule_registry.cc:321:36: error: reference to local binding 'lhsDim' declared in enclosing lambda expression
builder.addFactor({lhsDim, addRhs ? rhsDim : kNullDim},
^
external/shardy/shardy/dialect/sdy/transforms/propagation/op_sharding_rule_registry.cc:302:22: note: 'lhsDim' declared here
for (auto [lhsDim, rhsDim, outDim] :
^
external/shardy/shardy/dialect/sdy/transforms/propagation/op_sharding_rule_registry.cc:321:53: error: reference to local binding 'rhsDim' declared in enclosing lambda expression
builder.addFactor({lhsDim, addRhs ? rhsDim : kNullDim},
^
external/shardy/shardy/dialect/sdy/transforms/propagation/op_sharding_rule_registry.cc:302:30: note: 'rhsDim' declared here
for (auto [lhsDim, rhsDim, outDim] :
^
external/shardy/shardy/dialect/sdy/transforms/propagation/op_sharding_rule_registry.cc:322:44: error: reference to local binding 'outDim' declared in enclosing lambda expression
addOut ? outDim : kNullDim, factorSize);
^
external/shardy/shardy/dialect/sdy/transforms/propagation/op_sharding_rule_registry.cc:302:38: note: 'outDim' declared here
for (auto [lhsDim, rhsDim, outDim] :
—
Reply to this email directly, view it on GitHub
<#14352 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/A7DIFHZKYKMLCXQM3AD2JQDZMA6HRAVCNFSM6AAAAABKGSDSNGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMRWGMZTEMJZGQ>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
After pulling to latest main (which includes all of your changes @tomnatan30), I hit the following error:
@jreiffers, seems to be related to something you added. If I comment out this BUILD file the compile passes successfully. Thanks everyone for your help! |
Hey Aleks,
This is no longer a Shardy failure, will let @jreiffers
<https://github.com/jreiffers> comment on this.
…On Mon, Jul 15, 2024 at 4:17 PM Aleks Knezevic ***@***.***> wrote:
After pulling to latest main (which includes all of your changes
@tomnatan30 <https://github.com/tomnatan30>), I hit the following error:
ERROR: /localdev/aknezevic/xla/xla/service/gpu/model/fuzztest/BUILD:11:28: no such package 'tools': BUILD file not found in any of the following directories. Add a BUILD file to a directory to mark it as a package.
- /localdev/aknezevic/xla/tools and referenced by '//xla/service/gpu/model/fuzztest:affine_grammar_source'
ERROR: Analysis of target '//xla/service/gpu/model/fuzztest:affine_grammar_source' failed; build aborted: Analysis failed
INFO: Elapsed time: 17.830s
INFO: 0 processes.
FAILED: Build did NOT complete successfully (168 packages loaded, 688 targets configured)
@jreiffers <https://github.com/jreiffers>, seems to be related to
something you added. If I comment out this BUILD file the compile passes
successfully. Thanks everyone for your help!
—
Reply to this email directly, view it on GitHub
<#14352 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/A7DIFHZNOZ4JDDUKSF2XXVLZMPRXXAVCNFSM6AAAAABKGSDSNGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMRYG42TGNJUGU>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
Hm, this is a target generated by googlefuzztest. The relevant line is here. I think the problem is that this tool is within fuzztest, not within XLA, so the path is wrong when we generate the target within XLA. I'll try to find someone who built the tooling. |
@jreiffers I have already sent a change to googlefuzztest that I hope will fix this. |
#15060 is the attempt to update the revision of google_fuzztest that contains the fix. Unfortunately it is blocked because at head, google_fuzztest is now also using Riegeli, and we don't necessarily want to pull in that dependency as well. |
I'm trying to build xla from source for CPU following the instructions here and it's failing with:
I'm on an ubuntu 20.04 machine running a TF docker as the guide recommends. I have bazel 6.5.0 and clang 17. Is this config supported? Anything better to use?
If I run build with
-s
I see that-iquote .
is included, andxla/service/gpu/runtime/gpublas_lt_matmul_thunk.h
is there. Here's the full subcommand output:The text was updated successfully, but these errors were encountered: