Groups
Conversations
All groups and messages
Send feedback to Google
Help
Training
Sign in
Groups
OpenXLA Discuss
Conversations
About
Groups keyboard shortcuts have been updated
Dismiss
See shortcuts
OpenXLA Discuss
1â30 of 150
Welcome to the OpenXLA Project community!Â
Here are other ways to get involved in the project:Â
Attend an
OpenXLA Community Meeting
Subscribe to the
OpenXLA Youtube Channel
Follow theÂ
OpenXLA Twitter
Contribute to OpenXLA on
GitHub
Mark all as read
Report group
0 selected
Steeve Morin
Jan 14
tf.aliasing vs jax.buffer_donor
Hi folks. Is there a reason to use tf.aliasing instead of the higher level jax.buffer_donor ? It
unread,
tf.aliasing vs jax.buffer_donor
Hi folks. Is there a reason to use tf.aliasing instead of the higher level jax.buffer_donor ? It
Jan 14
Sandeep Dasgupta
,Â
Jan Pfeifer
6
Jan 14
HLO Componentization: Progress update
Hello Everybody, This is a follow-up to the previous announcement about the deprecation of certain
unread,
HLO Componentization: Progress update
Hello Everybody, This is a follow-up to the previous announcement about the deprecation of certain
Jan 14
Aditi Joshi
Jan 14
Updated invitation: OpenXLA Community Meeting - Monthly @ Tue Jan 14, 2025 9am - 9:30am (PST) (
[email protected]
)
OpenXLA Community Meeting - Monthly Join with Google Meet â OpenXLA Community Meeting Notes This
unread,
Updated invitation: OpenXLA Community Meeting - Monthly @ Tue Jan 14, 2025 9am - 9:30am (PST) (
[email protected]
)
OpenXLA Community Meeting - Monthly Join with Google Meet â OpenXLA Community Meeting Notes This
Jan 14
Aditi Joshi
8
Jan 14
July 16, Tues, 9-10 am PT OpenXLA Monthly Community meeting
Hi folks, Friendly reminder that tomorrow is our very first OpenXLA Community meeting of the 2025 (
unread,
July 16, Tues, 9-10 am PT OpenXLA Monthly Community meeting
Hi folks, Friendly reminder that tomorrow is our very first OpenXLA Community meeting of the 2025 (
Jan 14
Jack Wolfard
12/23/24
Timeline for upgrading Bazel 6.5.0 to a newer LTS
What is the timeline for OpenXLA projects to upgrade from Bazel 6.5.0 to a newer LTS like 8.0.0?
unread,
Timeline for upgrading Bazel 6.5.0 to a newer LTS
What is the timeline for OpenXLA projects to upgrade from Bazel 6.5.0 to a newer LTS like 8.0.0?
12/23/24
Sayce Falk
12/20/24
Updated invitation: OpenXLA Community Meeting - Monthly @ Monthly from 9am to 10am on the third Tuesday from Tue Sep 17 to Mon Jan 20, 2025 (PDT) (
[email protected]
)
OpenXLA Community Meeting - Monthly Join with Google Meet â OpenXLA Community Meeting Notes This
unread,
Updated invitation: OpenXLA Community Meeting - Monthly @ Monthly from 9am to 10am on the third Tuesday from Tue Sep 17 to Mon Jan 20, 2025 (PDT) (
[email protected]
)
OpenXLA Community Meeting - Monthly Join with Google Meet â OpenXLA Community Meeting Notes This
12/20/24
Cyril Bortolato
12/12/24
Updated invitation: OpenXLA Community Meeting - Monthly @ Tue Jan 14, 2025 9am - 9:45am (PST) (
[email protected]
)
OpenXLA Community Meeting - Monthly Join with Google Meet â OpenXLA Community Meeting Notes This
unread,
Updated invitation: OpenXLA Community Meeting - Monthly @ Tue Jan 14, 2025 9am - 9:45am (PST) (
[email protected]
)
OpenXLA Community Meeting - Monthly Join with Google Meet â OpenXLA Community Meeting Notes This
12/12/24
Muneeb Anwar
12/5/24
Inserting custom-calls inside HLOComputation for customized runtime instrumentation
Hi, I'm trying to rewrite HLOModule in order to insert custom-calls inside each HLOComputation.
unread,
Inserting custom-calls inside HLOComputation for customized runtime instrumentation
Hi, I'm trying to rewrite HLOModule in order to insert custom-calls inside each HLOComputation.
12/5/24
Ovidiu Marcu
, â¦Â
Bart Chrzaszcz
6
12/4/24
Shardy with PJRT
Hello, Thank you for your time and feedback. I am looking at the idea to have a runtime/middleware
unread,
Shardy with PJRT
Hello, Thank you for your time and feedback. I am looking at the idea to have a runtime/middleware
12/4/24
Cyril Bortolato
12/3/24
Updated invitation: OpenXLA Community Meeting - Monthly @ Tue Dec 17, 2024 9am - 10am (PST) (
[email protected]
)
OpenXLA Community Meeting - Monthly Join with Google Meet â OpenXLA Community Meeting Notes This
unread,
Updated invitation: OpenXLA Community Meeting - Monthly @ Tue Dec 17, 2024 9am - 10am (PST) (
[email protected]
)
OpenXLA Community Meeting - Monthly Join with Google Meet â OpenXLA Community Meeting Notes This
12/3/24
James Rubin
2
11/14/24
Tune in virtually to the OpenXLA Dev Lab tomorrow!
Hey Everyone, I forgot to include the time slots for the sessions mentioned above, so here they are:
unread,
Tune in virtually to the OpenXLA Dev Lab tomorrow!
Hey Everyone, I forgot to include the time slots for the sessions mentioned above, so here they are:
11/14/24
Ben Huntsman
, â¦Â
Peter Hawkins
5
11/12/24
Support for IBM Power CPU backend
One thing that seems plausible to me is that we could allow the user to provide an external or system
unread,
Support for IBM Power CPU backend
One thing that seems plausible to me is that we could allow the user to provide an external or system
11/12/24
Devesh Ratho
11/5/24
Layout API propagation across PJRT C-APIs
Hello everyone, I'm trying to understand the propagation of Layout information across PJRT C-APIs
unread,
Layout API propagation across PJRT C-APIs
Hello everyone, I'm trying to understand the propagation of Layout information across PJRT C-APIs
11/5/24
Gunhyun Park
10/28/24
Heads up: CHLO Experimental Ragged Ops
Hi everyone, We are working on a pull request to add experimental ragged ops to HLO, using optionally
unread,
Heads up: CHLO Experimental Ragged Ops
Hi everyone, We are working on a pull request to add experimental ragged ops to HLO, using optionally
10/28/24
Rachel Han
10/15/24
[RFC] Add result accuracy to transcendental unary ops
Hi everyone, I posted the RFC on adding support for selecting implementation of transcendental unary
unread,
[RFC] Add result accuracy to transcendental unary ops
Hi everyone, I posted the RFC on adding support for selecting implementation of transcendental unary
10/15/24
Aditi Joshi
10/14/24
Updated invitation: OpenXLA Community Meeting - Monthly @ Tue Oct 15, 2024 9am - 10am (PDT) (
[email protected]
)
OpenXLA Community Meeting - Monthly Join with Google Meet â OpenXLA Community Meeting Notes This
unread,
Updated invitation: OpenXLA Community Meeting - Monthly @ Tue Oct 15, 2024 9am - 10am (PDT) (
[email protected]
)
OpenXLA Community Meeting - Monthly Join with Google Meet â OpenXLA Community Meeting Notes This
10/14/24
James Rubin
10/11/24
Join Us for the 2nd OpenXLA Dev Lab
Hello OpenXLA Community! We're excited to invite you to our 2nd OpenXLA Dev Lab on November 14,
unread,
Join Us for the 2nd OpenXLA Dev Lab
Hello OpenXLA Community! We're excited to invite you to our 2nd OpenXLA Dev Lab on November 14,
10/11/24
Sergey Kozub CH
10/9/24
[RFC] Adding MX (microscaling) floating point types to StableHLO and XLA
Hello, everyone I posted RFCs for introducing MX primitive types to StableHLO and XLA: StableHLO RFC#
unread,
[RFC] Adding MX (microscaling) floating point types to StableHLO and XLA
Hello, everyone I posted RFCs for introducing MX primitive types to StableHLO and XLA: StableHLO RFC#
10/9/24
José Valim
9/28/24
Long term support for infeed/outfeed
Hi folks, Thank you so much for the continuous work on XLA. Without it, we would not be able to bring
unread,
Long term support for infeed/outfeed
Hi folks, Thank you so much for the continuous work on XLA. Without it, we would not be able to bring
9/28/24
Joey Ye
,Â
Eugene Zhulenev
2
9/27/24
Why was MLIR based CPU backend gone?
Maintenance. It was never fully implemented (launched by default) and was too much work to keep
unread,
Why was MLIR based CPU backend gone?
Maintenance. It was never fully implemented (launched by default) and was too much work to keep
9/27/24
Muneeb Anwar
, â¦Â
Eugene Zhulenev
5
9/27/24
Enabling hlo op profiling for cpu
To make this work you'd need to make changes to xla/backends/cpu/runtime/thunk_executor.h. This
unread,
Enabling hlo op profiling for cpu
To make this work you'd need to make changes to xla/backends/cpu/runtime/thunk_executor.h. This
9/27/24
Perry Gibson
, â¦Â
Jan Pfeifer
14
9/21/24
StableHLO and KV-cache
Thanks for the more specific reply Peter. You mentioned "you tell XLA you'd like to alias
unread,
StableHLO and KV-cache
Thanks for the more specific reply Peter. You mentioned "you tell XLA you'd like to alias
9/21/24
Cyril Bortolato
9/19/24
Updated invitation: OpenXLA Community Meeting - Monthly @ Tue Sep 24, 2024 9am - 9:45am (PDT) (
[email protected]
)
OpenXLA Community Meeting - Monthly Join with Google Meet â OpenXLA Community Meeting Notes This
unread,
Updated invitation: OpenXLA Community Meeting - Monthly @ Tue Sep 24, 2024 9am - 9:45am (PDT) (
[email protected]
)
OpenXLA Community Meeting - Monthly Join with Google Meet â OpenXLA Community Meeting Notes This
9/19/24
Aditi Joshi
2
9/17/24
Sept 18 Google reception after Day 1 of PyTorch conference from 6:30-9 pm
Friendly reminder to RSVP to the Google reception for OpenXLA this Wed, Sept 18, from 6:30-9 pm. You
unread,
Sept 18 Google reception after Day 1 of PyTorch conference from 6:30-9 pm
Friendly reminder to RSVP to the Google reception for OpenXLA this Wed, Sept 18, from 6:30-9 pm. You
9/17/24
Maksym Riabov
, â¦Â
Jan Pfeifer
4
9/14/24
Creating a scientific sim on OpenXLA
Let me offer yet another approach: there is the XlaBuilder library (see description of the operations
unread,
Creating a scientific sim on OpenXLA
Let me offer yet another approach: there is the XlaBuilder library (see description of the operations
9/14/24
Pushkar Ratnalikar
, â¦Â
xq Dan
8
9/13/24
Addition of more high-level Ops to StableHLO
Hi allï¼ I'd like to participate in the discussion on the extensibility issues of StableHLO. I
unread,
Addition of more high-level Ops to StableHLO
Hi allï¼ I'd like to participate in the discussion on the extensibility issues of StableHLO. I
9/13/24
Steeve Morin
, â¦Â
George Karpenkov
6
9/12/24
Very bad performance when xla_gpu_enable_triton_gemm=true on MI210/MI300X
> Quick side question: we usually run compilations in parallel (in different threads), does that
unread,
Very bad performance when xla_gpu_enable_triton_gemm=true on MI210/MI300X
> Quick side question: we usually run compilations in parallel (in different threads), does that
9/12/24
Alexander Pivovarov
9/10/24
Request for Review: PR-16585 Adding float8_e4m3 Support
Hello everyone, I am seeking help with the review of PR-16585 that introduces support for the
unread,
Request for Review: PR-16585 Adding float8_e4m3 Support
Hello everyone, I am seeking help with the review of PR-16585 that introduces support for the
9/10/24
Cyril Bortolato
9/10/24
Updated invitation: OpenXLA Community Meeting - Monthly @ Tue Sep 17, 2024 9am - 10am (PDT) (
[email protected]
)
OpenXLA Community Meeting - Monthly Join with Google Meet â OpenXLA Community Meeting Notes This
unread,
Updated invitation: OpenXLA Community Meeting - Monthly @ Tue Sep 17, 2024 9am - 10am (PDT) (
[email protected]
)
OpenXLA Community Meeting - Monthly Join with Google Meet â OpenXLA Community Meeting Notes This
9/10/24
Vishnu Ram OV
9/10/24
Updated invitation: OpenXLA Community Meeting - Monthly @ Monthly from 7pm to 8pm on the third Tuesday from Tue Jul 16 to Tue Sep 17 (GMT+3) (
[email protected]
)
OpenXLA Community Meeting - Monthly Join with Google Meet â OpenXLA Community Meeting Notes This
unread,
Updated invitation: OpenXLA Community Meeting - Monthly @ Monthly from 7pm to 8pm on the third Tuesday from Tue Jul 16 to Tue Sep 17 (GMT+3) (
[email protected]
)
OpenXLA Community Meeting - Monthly Join with Google Meet â OpenXLA Community Meeting Notes This
9/10/24