Open jwknaup opened 2 months ago
Seeing the same issue on a Jetson Orin AGX with JP6. I think it should be fixable with the patch from https://github.com/tensorflow/tensorflow/issues/62490 (applied to xla -> tsl -> absl
). I'm building with the patch now, I'll comment again if this fixes the issue.
FYI the above workaround works as a one-off; it would be nice to get the fix upstreamed once https://github.com/abseil/abseil-cpp/issues/1665 is resolved.
Thanks Kyle! Worked for me as well
@kylestach Hi, I am facing the same issue here with the Jetson Orin Nano. Did you figure out ?
Note that this has been fixed upstream in absl: https://github.com/abseil/abseil-cpp/pull/1732. Maybe we can bump the vendored copy of absl before the next release?
There is already a PR to add this patch in XLA: https://github.com/openxla/xla/pull/15687 So it should be fixed soon.
Did you link the wrong PR?
Updated my previous comment.
In that case can we bump the vendored copy of XLA? Or just stop vendoring things so it's no longer a JAX bug?
JAX seem to update the XLA commit everyday or so. I looked, current JAX upstream point to an XLA that has the patch. So taking today/tomorrow JAX nightly or building upstream should get you the fix.
If you try, can you confirm the fix works for you?
it is still not building... main branch
from external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc:18:
external/com_google_absl/absl/log/check.h:136: warning: "QCHECK_GE" redefined
136 | #define QCHECK_GE(val1, val2) \
|
In file included from external/tsl/tsl/platform/logging.h:26,
from external/xla/xla/stream_executor/device_memory.h:34,
from external/xla/xla/stream_executor/gpu/gpu_semaphore.h:23,
from external/xla/xla/stream_executor/cuda/delay_kernel.h:20,
from external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc:18:
external/tsl/tsl/platform/default/logging.h:543: note: this is the location of the previous definition
543 | #define QCHECK_GE(x, y) CHECK_GE(x, y)
|
In file included from external/xla/xla/stream_executor/stream.h:30,
from external/xla/xla/stream_executor/stream_executor.h:50,
from external/xla/xla/stream_executor/gpu/gpu_semaphore.h:25,
from external/xla/xla/stream_executor/cuda/delay_kernel.h:20,
from external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc:18:
external/com_google_absl/absl/log/check.h:138: warning: "QCHECK_GT" redefined
138 | #define QCHECK_GT(val1, val2) \
|
In file included from external/tsl/tsl/platform/logging.h:26,
from external/xla/xla/stream_executor/device_memory.h:34,
from external/xla/xla/stream_executor/gpu/gpu_semaphore.h:23,
from external/xla/xla/stream_executor/cuda/delay_kernel.h:20,
from external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc:18:
external/tsl/tsl/platform/default/logging.h:544: note: this is the location of the previous definition
544 | #define QCHECK_GT(x, y) CHECK_GT(x, y)
|
In file included from external/xla/xla/stream_executor/stream.h:30,
from external/xla/xla/stream_executor/stream_executor.h:50,
from external/xla/xla/stream_executor/gpu/gpu_semaphore.h:25,
from external/xla/xla/stream_executor/cuda/delay_kernel.h:20,
from external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc:18:
external/com_google_absl/absl/log/check.h:140: warning: "DCHECK_EQ" redefined
140 | #define DCHECK_EQ(val1, val2) \
|
In file included from external/tsl/tsl/platform/logging.h:26,
from external/xla/xla/stream_executor/device_memory.h:34,
from external/xla/xla/stream_executor/gpu/gpu_semaphore.h:23,
from external/xla/xla/stream_executor/cuda/delay_kernel.h:20,
from external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc:18:
external/tsl/tsl/platform/default/logging.h:527: note: this is the location of the previous definition
527 | #define DCHECK_EQ(x, y) _TF_DCHECK_NOP(x, y)
|
In file included from external/xla/xla/stream_executor/stream.h:30,
from external/xla/xla/stream_executor/stream_executor.h:50,
from external/xla/xla/stream_executor/gpu/gpu_semaphore.h:25,
from external/xla/xla/stream_executor/cuda/delay_kernel.h:20,
from external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc:18:
external/com_google_absl/absl/log/check.h:142: warning: "DCHECK_NE" redefined
142 | #define DCHECK_NE(val1, val2) \
|
In file included from external/tsl/tsl/platform/logging.h:26,
from external/xla/xla/stream_executor/device_memory.h:34,
from external/xla/xla/stream_executor/gpu/gpu_semaphore.h:23,
from external/xla/xla/stream_executor/cuda/delay_kernel.h:20,
from external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc:18:
external/tsl/tsl/platform/default/logging.h:528: note: this is the location of the previous definition
528 | #define DCHECK_NE(x, y) _TF_DCHECK_NOP(x, y)
|
In file included from external/xla/xla/stream_executor/stream.h:30,
from external/xla/xla/stream_executor/stream_executor.h:50,
from external/xla/xla/stream_executor/gpu/gpu_semaphore.h:25,
from external/xla/xla/stream_executor/cuda/delay_kernel.h:20,
from external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc:18:
external/com_google_absl/absl/log/check.h:144: warning: "DCHECK_LE" redefined
144 | #define DCHECK_LE(val1, val2) \
|
In file included from external/tsl/tsl/platform/logging.h:26,
from external/xla/xla/stream_executor/device_memory.h:34,
from external/xla/xla/stream_executor/gpu/gpu_semaphore.h:23,
from external/xla/xla/stream_executor/cuda/delay_kernel.h:20,
from external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc:18:
external/tsl/tsl/platform/default/logging.h:529: note: this is the location of the previous definition
529 | #define DCHECK_LE(x, y) _TF_DCHECK_NOP(x, y)
|
In file included from external/xla/xla/stream_executor/stream.h:30,
from external/xla/xla/stream_executor/stream_executor.h:50,
from external/xla/xla/stream_executor/gpu/gpu_semaphore.h:25,
from external/xla/xla/stream_executor/cuda/delay_kernel.h:20,
from external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc:18:
external/com_google_absl/absl/log/check.h:146: warning: "DCHECK_LT" redefined
146 | #define DCHECK_LT(val1, val2) \
|
In file included from external/tsl/tsl/platform/logging.h:26,
from external/xla/xla/stream_executor/device_memory.h:34,
from external/xla/xla/stream_executor/gpu/gpu_semaphore.h:23,
from external/xla/xla/stream_executor/cuda/delay_kernel.h:20,
from external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc:18:
external/tsl/tsl/platform/default/logging.h:530: note: this is the location of the previous definition
530 | #define DCHECK_LT(x, y) _TF_DCHECK_NOP(x, y)
|
In file included from external/xla/xla/stream_executor/stream.h:30,
from external/xla/xla/stream_executor/stream_executor.h:50,
from external/xla/xla/stream_executor/gpu/gpu_semaphore.h:25,
from external/xla/xla/stream_executor/cuda/delay_kernel.h:20,
from external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc:18:
external/com_google_absl/absl/log/check.h:148: warning: "DCHECK_GE" redefined
148 | #define DCHECK_GE(val1, val2) \
|
In file included from external/tsl/tsl/platform/logging.h:26,
from external/xla/xla/stream_executor/device_memory.h:34,
from external/xla/xla/stream_executor/gpu/gpu_semaphore.h:23,
from external/xla/xla/stream_executor/cuda/delay_kernel.h:20,
from external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc:18:
external/tsl/tsl/platform/default/logging.h:531: note: this is the location of the previous definition
531 | #define DCHECK_GE(x, y) _TF_DCHECK_NOP(x, y)
|
In file included from external/xla/xla/stream_executor/stream.h:30,
from external/xla/xla/stream_executor/stream_executor.h:50,
from external/xla/xla/stream_executor/gpu/gpu_semaphore.h:25,
from external/xla/xla/stream_executor/cuda/delay_kernel.h:20,
from external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc:18:
external/com_google_absl/absl/log/check.h:150: warning: "DCHECK_GT" redefined
150 | #define DCHECK_GT(val1, val2) \
|
In file included from external/tsl/tsl/platform/logging.h:26,
from external/xla/xla/stream_executor/device_memory.h:34,
from external/xla/xla/stream_executor/gpu/gpu_semaphore.h:23,
from external/xla/xla/stream_executor/cuda/delay_kernel.h:20,
from external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc:18:
external/tsl/tsl/platform/default/logging.h:532: note: this is the location of the previous definition
532 | #define DCHECK_GT(x, y) _TF_DCHECK_NOP(x, y)
|
external/com_google_absl/absl/status/status.h(796): warning #2810-D: ignoring return value type with "nodiscard" attribute
*this = std::move(new_status);
^
Remark: The warnings can be suppressed with "-diag-suppress <warning-number>"
external/com_google_absl/absl/status/internal/statusor_internal.h(240): warning #2810-D: ignoring return value type with "nodiscard" attribute
status_ = OkStatus();
^
external/com_google_absl/absl/status/internal/statusor_internal.h(247): warning #2810-D: ignoring return value type with "nodiscard" attribute
status_ = static_cast<absl::Status>(std::forward<U>(v));
^
external/xla/xla/stream_executor/stream_executor.h(68): warning #940-D: missing return statement at end of non-void function "stream_executor::MemoryTypeString"
}
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(44): error: identifier "__Poly8x8_t" is undefined
typedef __Poly8x8_t poly8x8_t;
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(45): error: identifier "__Poly16x4_t" is undefined
typedef __Poly16x4_t poly16x4_t;
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(58): error: identifier "__Poly8x16_t" is undefined
typedef __Poly8x16_t poly8x16_t;
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(59): error: identifier "__Poly16x8_t" is undefined
typedef __Poly16x8_t poly16x8_t;
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(60): error: identifier "__Poly64x2_t" is undefined
typedef __Poly64x2_t poly64x2_t;
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(61): error: identifier "__Poly64x1_t" is undefined
typedef __Poly64x1_t poly64x1_t;
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(67): error: identifier "__Poly8_t" is undefined
typedef __Poly8_t poly8_t;
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(68): error: identifier "__Poly16_t" is undefined
typedef __Poly16_t poly16_t;
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(69): error: identifier "__Poly64_t" is undefined
typedef __Poly64_t poly64_t;
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(70): error: identifier "__Poly128_t" is undefined
typedef __Poly128_t poly128_t;
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(828): error: identifier "__builtin_aarch64_saddlv8qi" is undefined
return (int16x8_t) __builtin_aarch64_saddlv8qi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(835): error: identifier "__builtin_aarch64_saddlv4hi" is undefined
return (int32x4_t) __builtin_aarch64_saddlv4hi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(842): error: identifier "__builtin_aarch64_saddlv2si" is undefined
return (int64x2_t) __builtin_aarch64_saddlv2si (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(849): error: identifier "__builtin_aarch64_uaddlv8qi" is undefined
return (uint16x8_t) __builtin_aarch64_uaddlv8qi ((int8x8_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(857): error: identifier "__builtin_aarch64_uaddlv4hi" is undefined
return (uint32x4_t) __builtin_aarch64_uaddlv4hi ((int16x4_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(865): error: identifier "__builtin_aarch64_uaddlv2si" is undefined
return (uint64x2_t) __builtin_aarch64_uaddlv2si ((int32x2_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(873): error: identifier "__builtin_aarch64_saddl2v16qi" is undefined
return (int16x8_t) __builtin_aarch64_saddl2v16qi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(880): error: identifier "__builtin_aarch64_saddl2v8hi" is undefined
return (int32x4_t) __builtin_aarch64_saddl2v8hi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(887): error: identifier "__builtin_aarch64_saddl2v4si" is undefined
return (int64x2_t) __builtin_aarch64_saddl2v4si (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(894): error: identifier "__builtin_aarch64_uaddl2v16qi" is undefined
return (uint16x8_t) __builtin_aarch64_uaddl2v16qi ((int8x16_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(902): error: identifier "__builtin_aarch64_uaddl2v8hi" is undefined
return (uint32x4_t) __builtin_aarch64_uaddl2v8hi ((int16x8_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(910): error: identifier "__builtin_aarch64_uaddl2v4si" is undefined
return (uint64x2_t) __builtin_aarch64_uaddl2v4si ((int32x4_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(918): error: identifier "__builtin_aarch64_saddwv8qi" is undefined
return (int16x8_t) __builtin_aarch64_saddwv8qi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(925): error: identifier "__builtin_aarch64_saddwv4hi" is undefined
return (int32x4_t) __builtin_aarch64_saddwv4hi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(932): error: identifier "__builtin_aarch64_saddwv2si" is undefined
return (int64x2_t) __builtin_aarch64_saddwv2si (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(939): error: identifier "__builtin_aarch64_uaddwv8qi" is undefined
return (uint16x8_t) __builtin_aarch64_uaddwv8qi ((int16x8_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(947): error: identifier "__builtin_aarch64_uaddwv4hi" is undefined
return (uint32x4_t) __builtin_aarch64_uaddwv4hi ((int32x4_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(955): error: identifier "__builtin_aarch64_uaddwv2si" is undefined
return (uint64x2_t) __builtin_aarch64_uaddwv2si ((int64x2_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(963): error: identifier "__builtin_aarch64_saddw2v16qi" is undefined
return (int16x8_t) __builtin_aarch64_saddw2v16qi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(970): error: identifier "__builtin_aarch64_saddw2v8hi" is undefined
return (int32x4_t) __builtin_aarch64_saddw2v8hi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(977): error: identifier "__builtin_aarch64_saddw2v4si" is undefined
return (int64x2_t) __builtin_aarch64_saddw2v4si (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(984): error: identifier "__builtin_aarch64_uaddw2v16qi" is undefined
return (uint16x8_t) __builtin_aarch64_uaddw2v16qi ((int16x8_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(992): error: identifier "__builtin_aarch64_uaddw2v8hi" is undefined
return (uint32x4_t) __builtin_aarch64_uaddw2v8hi ((int32x4_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1000): error: identifier "__builtin_aarch64_uaddw2v4si" is undefined
return (uint64x2_t) __builtin_aarch64_uaddw2v4si ((int64x2_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1008): error: identifier "__builtin_aarch64_shaddv8qi" is undefined
return (int8x8_t) __builtin_aarch64_shaddv8qi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1015): error: identifier "__builtin_aarch64_shaddv4hi" is undefined
return (int16x4_t) __builtin_aarch64_shaddv4hi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1022): error: identifier "__builtin_aarch64_shaddv2si" is undefined
return (int32x2_t) __builtin_aarch64_shaddv2si (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1029): error: identifier "__builtin_aarch64_uhaddv8qi" is undefined
return (uint8x8_t) __builtin_aarch64_uhaddv8qi ((int8x8_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1037): error: identifier "__builtin_aarch64_uhaddv4hi" is undefined
return (uint16x4_t) __builtin_aarch64_uhaddv4hi ((int16x4_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1045): error: identifier "__builtin_aarch64_uhaddv2si" is undefined
return (uint32x2_t) __builtin_aarch64_uhaddv2si ((int32x2_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1053): error: identifier "__builtin_aarch64_shaddv16qi" is undefined
return (int8x16_t) __builtin_aarch64_shaddv16qi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1060): error: identifier "__builtin_aarch64_shaddv8hi" is undefined
return (int16x8_t) __builtin_aarch64_shaddv8hi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1067): error: identifier "__builtin_aarch64_shaddv4si" is undefined
return (int32x4_t) __builtin_aarch64_shaddv4si (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1074): error: identifier "__builtin_aarch64_uhaddv16qi" is undefined
return (uint8x16_t) __builtin_aarch64_uhaddv16qi ((int8x16_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1082): error: identifier "__builtin_aarch64_uhaddv8hi" is undefined
return (uint16x8_t) __builtin_aarch64_uhaddv8hi ((int16x8_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1090): error: identifier "__builtin_aarch64_uhaddv4si" is undefined
return (uint32x4_t) __builtin_aarch64_uhaddv4si ((int32x4_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1098): error: identifier "__builtin_aarch64_srhaddv8qi" is undefined
return (int8x8_t) __builtin_aarch64_srhaddv8qi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1105): error: identifier "__builtin_aarch64_srhaddv4hi" is undefined
return (int16x4_t) __builtin_aarch64_srhaddv4hi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1112): error: identifier "__builtin_aarch64_srhaddv2si" is undefined
return (int32x2_t) __builtin_aarch64_srhaddv2si (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1119): error: identifier "__builtin_aarch64_urhaddv8qi" is undefined
return (uint8x8_t) __builtin_aarch64_urhaddv8qi ((int8x8_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1127): error: identifier "__builtin_aarch64_urhaddv4hi" is undefined
return (uint16x4_t) __builtin_aarch64_urhaddv4hi ((int16x4_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1135): error: identifier "__builtin_aarch64_urhaddv2si" is undefined
return (uint32x2_t) __builtin_aarch64_urhaddv2si ((int32x2_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1143): error: identifier "__builtin_aarch64_srhaddv16qi" is undefined
return (int8x16_t) __builtin_aarch64_srhaddv16qi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1150): error: identifier "__builtin_aarch64_srhaddv8hi" is undefined
return (int16x8_t) __builtin_aarch64_srhaddv8hi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1157): error: identifier "__builtin_aarch64_srhaddv4si" is undefined
return (int32x4_t) __builtin_aarch64_srhaddv4si (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1164): error: identifier "__builtin_aarch64_urhaddv16qi" is undefined
return (uint8x16_t) __builtin_aarch64_urhaddv16qi ((int8x16_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1172): error: identifier "__builtin_aarch64_urhaddv8hi" is undefined
return (uint16x8_t) __builtin_aarch64_urhaddv8hi ((int16x8_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1180): error: identifier "__builtin_aarch64_urhaddv4si" is undefined
return (uint32x4_t) __builtin_aarch64_urhaddv4si ((int32x4_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1188): error: identifier "__builtin_aarch64_addhnv8hi" is undefined
return (int8x8_t) __builtin_aarch64_addhnv8hi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1195): error: identifier "__builtin_aarch64_addhnv4si" is undefined
return (int16x4_t) __builtin_aarch64_addhnv4si (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1202): error: identifier "__builtin_aarch64_addhnv2di" is undefined
return (int32x2_t) __builtin_aarch64_addhnv2di (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1209): error: identifier "__builtin_aarch64_addhnv8hi" is undefined
return (uint8x8_t) __builtin_aarch64_addhnv8hi ((int16x8_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1217): error: identifier "__builtin_aarch64_addhnv4si" is undefined
return (uint16x4_t) __builtin_aarch64_addhnv4si ((int32x4_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1225): error: identifier "__builtin_aarch64_addhnv2di" is undefined
return (uint32x2_t) __builtin_aarch64_addhnv2di ((int64x2_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1233): error: identifier "__builtin_aarch64_raddhnv8hi" is undefined
return (int8x8_t) __builtin_aarch64_raddhnv8hi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1240): error: identifier "__builtin_aarch64_raddhnv4si" is undefined
return (int16x4_t) __builtin_aarch64_raddhnv4si (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1247): error: identifier "__builtin_aarch64_raddhnv2di" is undefined
return (int32x2_t) __builtin_aarch64_raddhnv2di (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1254): error: identifier "__builtin_aarch64_raddhnv8hi" is undefined
return (uint8x8_t) __builtin_aarch64_raddhnv8hi ((int16x8_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1262): error: identifier "__builtin_aarch64_raddhnv4si" is undefined
return (uint16x4_t) __builtin_aarch64_raddhnv4si ((int32x4_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1270): error: identifier "__builtin_aarch64_raddhnv2di" is undefined
return (uint32x2_t) __builtin_aarch64_raddhnv2di ((int64x2_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1278): error: identifier "__builtin_aarch64_addhn2v8hi" is undefined
return (int8x16_t) __builtin_aarch64_addhn2v8hi (__a, __b, __c);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1285): error: identifier "__builtin_aarch64_addhn2v4si" is undefined
return (int16x8_t) __builtin_aarch64_addhn2v4si (__a, __b, __c);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1292): error: identifier "__builtin_aarch64_addhn2v2di" is undefined
return (int32x4_t) __builtin_aarch64_addhn2v2di (__a, __b, __c);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1299): error: identifier "__builtin_aarch64_addhn2v8hi" is undefined
return (uint8x16_t) __builtin_aarch64_addhn2v8hi ((int8x8_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1308): error: identifier "__builtin_aarch64_addhn2v4si" is undefined
return (uint16x8_t) __builtin_aarch64_addhn2v4si ((int16x4_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1317): error: identifier "__builtin_aarch64_addhn2v2di" is undefined
return (uint32x4_t) __builtin_aarch64_addhn2v2di ((int32x2_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1326): error: identifier "__builtin_aarch64_raddhn2v8hi" is undefined
return (int8x16_t) __builtin_aarch64_raddhn2v8hi (__a, __b, __c);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1333): error: identifier "__builtin_aarch64_raddhn2v4si" is undefined
return (int16x8_t) __builtin_aarch64_raddhn2v4si (__a, __b, __c);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1340): error: identifier "__builtin_aarch64_raddhn2v2di" is undefined
return (int32x4_t) __builtin_aarch64_raddhn2v2di (__a, __b, __c);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1347): error: identifier "__builtin_aarch64_raddhn2v8hi" is undefined
return (uint8x16_t) __builtin_aarch64_raddhn2v8hi ((int8x8_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1356): error: identifier "__builtin_aarch64_raddhn2v4si" is undefined
return (uint16x8_t) __builtin_aarch64_raddhn2v4si ((int16x4_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1365): error: identifier "__builtin_aarch64_raddhn2v2di" is undefined
return (uint32x4_t) __builtin_aarch64_raddhn2v2di ((int32x2_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1458): error: identifier "__builtin_aarch64_pmulv8qi" is undefined
return (poly8x8_t) __builtin_aarch64_pmulv8qi ((int8x8_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(1522): error: identifier "__builtin_aarch64_pmulv16qi" is undefined
return (poly8x16_t) __builtin_aarch64_pmulv16qi ((int8x16_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(2230): error: identifier "__builtin_aarch64_ssublv8qi" is undefined
return (int16x8_t) __builtin_aarch64_ssublv8qi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(2237): error: identifier "__builtin_aarch64_ssublv4hi" is undefined
return (int32x4_t) __builtin_aarch64_ssublv4hi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(2244): error: identifier "__builtin_aarch64_ssublv2si" is undefined
return (int64x2_t) __builtin_aarch64_ssublv2si (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(2251): error: identifier "__builtin_aarch64_usublv8qi" is undefined
return (uint16x8_t) __builtin_aarch64_usublv8qi ((int8x8_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(2259): error: identifier "__builtin_aarch64_usublv4hi" is undefined
return (uint32x4_t) __builtin_aarch64_usublv4hi ((int16x4_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(2267): error: identifier "__builtin_aarch64_usublv2si" is undefined
return (uint64x2_t) __builtin_aarch64_usublv2si ((int32x2_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(2275): error: identifier "__builtin_aarch64_ssubl2v16qi" is undefined
return (int16x8_t) __builtin_aarch64_ssubl2v16qi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(2282): error: identifier "__builtin_aarch64_ssubl2v8hi" is undefined
return (int32x4_t) __builtin_aarch64_ssubl2v8hi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(2289): error: identifier "__builtin_aarch64_ssubl2v4si" is undefined
return (int64x2_t) __builtin_aarch64_ssubl2v4si (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(2296): error: identifier "__builtin_aarch64_usubl2v16qi" is undefined
return (uint16x8_t) __builtin_aarch64_usubl2v16qi ((int8x16_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(2304): error: identifier "__builtin_aarch64_usubl2v8hi" is undefined
return (uint32x4_t) __builtin_aarch64_usubl2v8hi ((int16x8_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(2312): error: identifier "__builtin_aarch64_usubl2v4si" is undefined
return (uint64x2_t) __builtin_aarch64_usubl2v4si ((int32x4_t) __a,
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(2320): error: identifier "__builtin_aarch64_ssubwv8qi" is undefined
return (int16x8_t) __builtin_aarch64_ssubwv8qi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(2327): error: identifier "__builtin_aarch64_ssubwv4hi" is undefined
return (int32x4_t) __builtin_aarch64_ssubwv4hi (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(2334): error: identifier "__builtin_aarch64_ssubwv2si" is undefined
return (int64x2_t) __builtin_aarch64_ssubwv2si (__a, __b);
^
/usr/lib/gcc/aarch64-linux-gnu/11/include/arm_neon.h(2341): error: identifier "__builtin_aarch64_usubwv8qi" is undefined
return (uint16x8_t) __builtin_aarch64_usubwv8qi ((int16x8_t) __a,
^
Error limit reached.
100 errors detected in the compilation of "external/xla/xla/stream_executor/cuda/delay_kernel_cuda.cu.cc".
Compilation terminated.
Target //jaxlib/tools:build_wheel failed to build
INFO: Elapsed time: 721.544s, Critical Path: 196.07s
INFO: 2716 processes: 20 internal, 2696 local.
FAILED: Build did NOT complete successfully
ERROR: Build failed. Not running target
Traceback (most recent call last):
File "/home/johnny/Projects/jax/build/build.py", line 749, in <module>
main()
File "/home/johnny/Projects/jax/build/build.py", line 700, in main
shell(build_cpu_wheel_command)
File "/home/johnny/Projects/jax/build/build.py", line 45, in shell
output = subprocess.check_output(cmd)
File "/usr/lib/python3.10/subprocess.py", line 421, in check_output
return run(*popenargs, stdout=PIPE, timeout=timeout, check=True,
File "/usr/lib/python3.10/subprocess.py", line 526, in run
raise CalledProcessError(retcode, process.args,
subprocess.CalledProcessError: Command '['./bazel-6.5.0-linux-arm64', 'run', '--verbose_failures=true', '--repo_env=LOCAL_CUDA_PATH=/usr/local/cuda-12.2', '--repo_env=LOCAL_CUDNN_PATH=/usr/lib/aarch64-linux-gnu', '//jaxlib/tools:build_wheel', '--', '--output_path=/home/johnny/Projects/jax/dist', '--jaxlib_git_hash=', '--cpu=aarch64']' returned non-zero exit status 1.
(py310) johnny@johnny-jetson:~/Projects/jax$
openxla/xla#15687
how to apply it?
@jwknaup do yo know how to apply this patch while it is building?
I believe you have to copy the files over and overwrite the existing ones in your local XLA code. It's been a while, though, so I don't remember what exactly that looked like or where they were. Sorry, I'm not much help. 😢
I believe you have to copy the files over and overwrite the existing ones in your local XLA code. It's been a while, though, so I don't remember what exactly that looked like or where they were. Sorry, I'm not much help. 😢
well, I will wait for next release... because they change a lot of cuda gpus
I've studied last versions and now I did some progress. There are a bug that clang no detect automatically version of cuda, so in case that put only path not works. All must be match. Also , be carefule with cudnn headers that are different.
the command is:
python3 build/build.py --enable_cuda --enable_nccl=False \
--cuda_compute_capabilities=sm_87 \
--cuda_version=12.2.0 --cudnn_version=8.9.4 \
--bazel_options=--repo_env=LOCAL_CUDA_PATH="/usr/local/cuda-12.2" \
--bazel_options=--repo_env=LOCAL_CUDNN_PATH="/usr/lib/aarch64-linux-gnu"
If you upgrade LLVM/clang first, that compiler error goes away (in my case with LLVM 18)
@jwknaup you have now the wheels on http://jetson.webredirect.org/jp6/cu126. Please if you get error install:
pip3 install --no_deps *.whl
Close the issue :)
Description
When I attempt to build jaxlib v0.4.30 on a Jetson Orin with the command
python build/build.py --enable_cuda --cuda_path /usr/local/cuda-12.2 --cudnn_path /usr/lib/aarch64-linux-gnu --cuda_version 12.2 --cudnn_version 8
I get the errorERROR: /home/nvidia/.cache/bazel/_bazel_nvidia/a5643b5cc286b9b13a96818003a4a7dd/external/xla/xla/stream_executor/gpu/BUILD:349:19: Compiling xla/stream_executor/gpu/gpu_timer_kernel_cuda.cu.cc failed: (Exit 4): crosstool_wrapper_driver_is_not_gcc failed: error executing command (from target @xla//xla/stream_executor/gpu:gpu_timer_kernel_cuda)
The full output is below
System info (python version, jaxlib version, accelerator, etc.)
I am using a Jetson Orin Nano with Tegra GPU running Jetpack 6
Cuda 12.2
cuDNN 8