This is an automated email from the ASF dual-hosted git repository.
github-bot pushed a change to branch nightly
in repository https://gitbox.apache.org/repos/asf/tvm.git
from d1ac1c0202 [KVCache] Fix the aux data syncing order of paged KV cache
(#16988)
add 1d4b9ea5c3 [UnitTest] Use
mawnja opened a new pull request, #16996:
URL: https://github.com/apache/tvm/pull/16996
[fixed to make TupleGetItem inherits the previous
span](https://github.com/apache/tvm/commit/0d14b3cb16c9e0a0087d2c2dc1b82cb6023d93a2)
--
This is an automated message from the Apache Git Service.
To
This is an automated email from the ASF dual-hosted git repository.
masahi pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git
The following commit(s) were added to refs/heads/main by this push:
new c2d14ae872 [Relax][Transform] Handle identical
masahi merged PR #16959:
URL: https://github.com/apache/tvm/pull/16959
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail:
cyx-6 opened a new pull request, #16995:
URL: https://github.com/apache/tvm/pull/16995
This PR supports the KVCache to decode from the forked sequence, and pop
trailing tokens over multiple blocks.
cc: @tqchen @MasterJH5574
--
This is an automated message from the Apache Git
Lunderberg commented on code in PR #16994:
URL: https://github.com/apache/tvm/pull/16994#discussion_r1599039598
##
python/tvm/_ffi/runtime_ctypes.py:
##
@@ -539,11 +539,25 @@ def total_global_memory(self):
Returns
---
total_global_memory : int or
Lunderberg commented on code in PR #16980:
URL: https://github.com/apache/tvm/pull/16980#discussion_r1599038038
##
src/runtime/cuda/cuda_device_api.cc:
##
@@ -142,6 +142,24 @@ class CUDADeviceAPI final : public DeviceAPI {
}
void FreeDataSpace(Device dev, void* ptr)
This is an automated email from the ASF dual-hosted git repository.
masahi pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git
The following commit(s) were added to refs/heads/main by this push:
new 5b5f8d0f77 [QoL][IR] Provide std::hash and
masahi merged PR #16909:
URL: https://github.com/apache/tvm/pull/16909
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail:
Lunderberg commented on PR #16978:
URL: https://github.com/apache/tvm/pull/16978#issuecomment-2108697140
Thank you, and conflict resolved!
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the
masahi commented on PR #16978:
URL: https://github.com/apache/tvm/pull/16978#issuecomment-2108674980
@Lunderberg Need to resolve the conflict.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to
This is an automated email from the ASF dual-hosted git repository.
masahi pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git
The following commit(s) were added to refs/heads/main by this push:
new 0dfc5f955e [Unity] Check for transpose and dynamic
masahi merged PR #16589:
URL: https://github.com/apache/tvm/pull/16589
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail:
masahi merged PR #16958:
URL: https://github.com/apache/tvm/pull/16958
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail:
This is an automated email from the ASF dual-hosted git repository.
masahi pushed a change to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git
from 29337449db [Cuda] Skip FreeDataSpace when CUDA driver is in
inconsistent state (#16980)
add eb242ec77b [DLight]
This is an automated email from the ASF dual-hosted git repository.
masahi pushed a change to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git
from fd820ade5f [Disco] Expose disco.Session.shutdown through the python
API (#16979)
add 29337449db [Cuda] Skip
masahi commented on code in PR #16980:
URL: https://github.com/apache/tvm/pull/16980#discussion_r1599001463
##
src/runtime/cuda/cuda_device_api.cc:
##
@@ -142,6 +142,24 @@ class CUDADeviceAPI final : public DeviceAPI {
}
void FreeDataSpace(Device dev, void* ptr) final {
masahi merged PR #16980:
URL: https://github.com/apache/tvm/pull/16980
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail:
This is an automated email from the ASF dual-hosted git repository.
masahi pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git
The following commit(s) were added to refs/heads/main by this push:
new fd820ade5f [Disco] Expose disco.Session.shutdown
masahi merged PR #16979:
URL: https://github.com/apache/tvm/pull/16979
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail:
This is an automated email from the ASF dual-hosted git repository.
masahi pushed a change to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git
from d1ac1c0202 [KVCache] Fix the aux data syncing order of paged KV cache
(#16988)
add 1d4b9ea5c3 [UnitTest] Use
masahi merged PR #16930:
URL: https://github.com/apache/tvm/pull/16930
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail:
masahi commented on code in PR #16994:
URL: https://github.com/apache/tvm/pull/16994#discussion_r1598985475
##
python/tvm/_ffi/runtime_ctypes.py:
##
@@ -539,11 +539,25 @@ def total_global_memory(self):
Returns
---
total_global_memory : int or None
Lunderberg closed pull request #16990: [Disco] Implement `num_workers` property
for `disco.Session`
URL: https://github.com/apache/tvm/pull/16990
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the
Lunderberg commented on PR #16990:
URL: https://github.com/apache/tvm/pull/16990#issuecomment-2108326695
Closing this PR, as it is identical to
https://github.com/apache/tvm/pull/16978. I think that means I have too many
open PRs, and any help in reviewing them would be appreciated.
--
Lunderberg opened a new pull request, #16994:
URL: https://github.com/apache/tvm/pull/16994
Prior to this commit, the total device memory could be queried through the
`DeviceAPI` interface, but the currently available device memory could not.
This functionality may be useful for
Lunderberg opened a new pull request, #16993:
URL: https://github.com/apache/tvm/pull/16993
The `disco.Session.scatter_from_worker0` function expects a `DRef` which an
`NDArray` on worker 0, and `NullOpt` on all other workers. Prior to this
commit, there was no method in the
Lunderberg commented on PR #16966:
URL: https://github.com/apache/tvm/pull/16966#issuecomment-2108122470
No problem, and thank you on the revisions!
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go
Lunderberg commented on code in PR #16966:
URL: https://github.com/apache/tvm/pull/16966#discussion_r1598726836
##
src/target/llvm/codegen_llvm.cc:
##
@@ -1768,11 +1774,17 @@ llvm::Value* CodeGenLLVM::VisitExpr_(const
BufferLoadNode* op) {
std::vector loads;
- auto
Lunderberg commented on code in PR #16966:
URL: https://github.com/apache/tvm/pull/16966#discussion_r1598722953
##
python/tvm/tir/buffer.py:
##
@@ -141,6 +141,57 @@ def vstore(self, begin, value):
begin = (begin,) if isinstance(begin, (int, PrimExpr)) else begin
Lunderberg commented on code in PR #16966:
URL: https://github.com/apache/tvm/pull/16966#discussion_r1598715370
##
src/tir/transforms/vectorize_loop.cc:
##
@@ -72,6 +72,126 @@ inline PrimExpr BroadcastTo(PrimExpr e, int lanes, bool
is_scalable) {
return Broadcast(e,
Lunderberg opened a new pull request, #16992:
URL: https://github.com/apache/tvm/pull/16992
Prior to this commit, using `disco.Session` methods to transfer `NDArray`
instances to workers could raise an exception if the `NDArray` is larger than
the buffer allocated by the OS for the
lhutton1 commented on code in PR #16966:
URL: https://github.com/apache/tvm/pull/16966#discussion_r1598578740
##
src/tir/transforms/vectorize_loop.cc:
##
@@ -72,6 +72,126 @@ inline PrimExpr BroadcastTo(PrimExpr e, int lanes, bool
is_scalable) {
return Broadcast(e,
Jupiterghy opened a new issue, #16991:
URL: https://github.com/apache/tvm/issues/16991
Executing a sequence of Relay transformations iteratively using a custom
script results in a segmentation fault (Segmentation fault (core dumped)). This
issue seems to be related to the quantity of
Lunderberg opened a new pull request, #16990:
URL: https://github.com/apache/tvm/pull/16990
Prior to this commit, while the `num_workers` argument was provided to the
`disco.Session` object, it could not be determined from an existing
`disco.Session` object. As a result, functions that
35 matches
Mail list logo