source stringclasses 13
values | gpu stringclasses 4
values | model stringclasses 13
values | kernel_family stringclasses 7
values | kernel_name stringclasses 328
values | dtype stringclasses 2
values | held_out bool 2
classes | M float64 1 16.4k ⌀ | N float64 8 248k ⌀ | K float64 2.05k 29.6k ⌀ | bs float64 1 1 ⌀ | seq float64 1 16.4k ⌀ | n_heads float64 8 64 ⌀ | head_dim float64 64 256 ⌀ | kv_heads float64 1 64 ⌀ | numel float64 64 531M ⌀ | op_type stringclasses 177
values | gpu_time_duration_ms float64 0 703 | launch_block_size float64 16 512 | launch_grid_size float64 1 1.04M | dram_bytes_sum float64 1.27 94.3B ⌀ | launch_registers_per_thread float64 16 255 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | mul | 0.00675 | 128 | 1,024 | 1,110,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::neg_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 9)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[l... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | neg | 0.00614 | 128 | 512 | 547,330 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | copy | void at::native::<unnamed>::CatArrayBatchedCopy<at::native::<unnamed>::OpaqueType<(unsigned int)2>, unsigned int, (int)4, (int)64, (int)64>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | other | 0.00861 | 512 | 528 | 1,060,000 | 25 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | mul | 0.00646 | 128 | 1,024 | 1,110,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<c10::BFloat16>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | residual | 0.00678 | 128 | 1,024 | 2,120,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 131,072 | mul | 0.00624 | 128 | 256 | 321,790 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::neg_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 9)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[l... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 65,536 | neg | 0.00621 | 128 | 128 | 154,110 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | copy | void at::native::<unnamed>::CatArrayBatchedCopy<at::native::<unnamed>::OpaqueType<(unsigned int)2>, unsigned int, (int)4, (int)64, (int)64>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | other | 0.00662 | 512 | 528 | 269,060 | 25 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 131,072 | mul | 0.00611 | 128 | 256 | 322,050 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<c10::BFloat16>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 131,072 | residual | 0.00634 | 128 | 256 | 551,170 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | copy | void at::native::<unnamed>::CatArrayBatchedCopy<at::native::<unnamed>::OpaqueType<(unsigned int)2>, unsigned int, (int)4, (int)64, (int)64>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | other | 0.00672 | 512 | 528 | 269,310 | 25 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | copy | void at::native::<unnamed>::CatArrayBatchedCopy<at::native::<unnamed>::OpaqueType<(unsigned int)2>, unsigned int, (int)4, (int)64, (int)64>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | other | 0.0063 | 512 | 528 | 269,310 | 25 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::AUnaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | mul | 0.00509 | 128 | 512 | 1,050,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::CUDAFunctor_add<c10::BFloat16>, std::array<char *, (unsigned long)3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | residual | 0.00474 | 128 | 512 | 2,100,000 | 34 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | cast | void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.00618 | 128 | 1,024 | 1,080,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, void at::native::<unnamed>::pow_tensor_scalar_kernel_impl<float, float>(at::TensorIteratorBase &, T2)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | rmsnorm | 0.0049 | 128 | 512 | 2,100,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | reduce | void at::native::reduce_kernel<(int)512, (int)1, at::native::ReduceOp<float, at::native::MeanOps<float, float, float, float>, unsigned int, float, (int)4, (int)4>>(T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 16,384 | rmsnorm | 0.01824 | 512 | 8 | 2,160,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | residual | 0.00358 | 128 | 1 | 5,890 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, at::native::rsqrt_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | rmsnorm | 0.00365 | 128 | 1 | 7,170 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | mul | 0.00582 | 128 | 2,048 | 2,110,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | cast | void at::native::vectorized_elementwise_kernel<(int)8, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | cast | 0.00477 | 128 | 512 | 2,100,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | mul | 0.00624 | 128 | 1,024 | 1,080,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::<unnamed>::silu_kernel(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 6)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | silu | 0.0063 | 128 | 1,600 | 3,290,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | mul | 0.00595 | 128 | 1,600 | 6,560,000 | 40 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::AUnaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | mul | 0.004 | 128 | 512 | 1,050,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::CUDAFunctor_add<c10::BFloat16>, std::array<char *, (unsigned long)3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | residual | 0.00502 | 128 | 512 | 2,100,000 | 34 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | cast | void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.00611 | 128 | 1,024 | 1,080,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, void at::native::<unnamed>::pow_tensor_scalar_kernel_impl<float, float>(at::TensorIteratorBase &, T2)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | rmsnorm | 0.00413 | 128 | 512 | 2,100,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | reduce | void at::native::reduce_kernel<(int)512, (int)1, at::native::ReduceOp<float, at::native::MeanOps<float, float, float, float>, unsigned int, float, (int)4, (int)4>>(T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 16,384 | rmsnorm | 0.01805 | 512 | 8 | 2,160,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | residual | 0.00362 | 128 | 1 | 5,890 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, at::native::rsqrt_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | rmsnorm | 0.00358 | 128 | 1 | 6,910 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | mul | 0.00582 | 128 | 2,048 | 2,110,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | cast | void at::native::vectorized_elementwise_kernel<(int)8, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | cast | 0.00419 | 128 | 512 | 2,100,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | mul | 0.0064 | 128 | 1,024 | 1,080,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | mul | 0.00659 | 128 | 1,024 | 1,110,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::neg_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 9)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[l... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | neg | 0.00611 | 128 | 512 | 547,580 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | copy | void at::native::<unnamed>::CatArrayBatchedCopy<at::native::<unnamed>::OpaqueType<(unsigned int)2>, unsigned int, (int)4, (int)64, (int)64>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | other | 0.00874 | 512 | 528 | 1,060,000 | 25 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | mul | 0.00659 | 128 | 1,024 | 1,110,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<c10::BFloat16>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | residual | 0.00669 | 128 | 1,024 | 2,120,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 131,072 | mul | 0.0063 | 128 | 256 | 322,300 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::neg_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 9)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[l... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 65,536 | neg | 0.00598 | 128 | 128 | 153,860 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | copy | void at::native::<unnamed>::CatArrayBatchedCopy<at::native::<unnamed>::OpaqueType<(unsigned int)2>, unsigned int, (int)4, (int)64, (int)64>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | other | 0.00637 | 512 | 528 | 269,570 | 25 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 131,072 | mul | 0.00605 | 128 | 256 | 321,790 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<c10::BFloat16>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 131,072 | residual | 0.00627 | 128 | 256 | 551,170 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | copy | void at::native::<unnamed>::CatArrayBatchedCopy<at::native::<unnamed>::OpaqueType<(unsigned int)2>, unsigned int, (int)4, (int)64, (int)64>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | other | 0.0064 | 512 | 528 | 269,570 | 25 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | copy | void at::native::<unnamed>::CatArrayBatchedCopy<at::native::<unnamed>::OpaqueType<(unsigned int)2>, unsigned int, (int)4, (int)64, (int)64>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | other | 0.00666 | 512 | 528 | 269,060 | 25 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::AUnaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | mul | 0.00432 | 128 | 512 | 1,050,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::CUDAFunctor_add<c10::BFloat16>, std::array<char *, (unsigned long)3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | residual | 0.00451 | 128 | 512 | 2,100,000 | 34 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | cast | void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.0064 | 128 | 1,024 | 1,080,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, void at::native::<unnamed>::pow_tensor_scalar_kernel_impl<float, float>(at::TensorIteratorBase &, T2)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | rmsnorm | 0.00426 | 128 | 512 | 2,100,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | reduce | void at::native::reduce_kernel<(int)512, (int)1, at::native::ReduceOp<float, at::native::MeanOps<float, float, float, float>, unsigned int, float, (int)4, (int)4>>(T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 16,384 | rmsnorm | 0.01894 | 512 | 8 | 2,160,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | residual | 0.00339 | 128 | 1 | 5,890 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, at::native::rsqrt_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | rmsnorm | 0.00365 | 128 | 1 | 7,170 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | mul | 0.00582 | 128 | 2,048 | 2,110,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | cast | void at::native::vectorized_elementwise_kernel<(int)8, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | cast | 0.0049 | 128 | 512 | 2,100,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | mul | 0.00643 | 128 | 1,024 | 1,080,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::<unnamed>::silu_kernel(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 6)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | silu | 0.00602 | 128 | 1,600 | 3,290,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | mul | 0.0065 | 128 | 1,600 | 6,560,000 | 40 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::AUnaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | mul | 0.00451 | 128 | 512 | 1,050,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::CUDAFunctor_add<c10::BFloat16>, std::array<char *, (unsigned long)3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | residual | 0.00422 | 128 | 512 | 2,100,000 | 34 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | cast | void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.00669 | 128 | 1,024 | 1,080,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, void at::native::<unnamed>::pow_tensor_scalar_kernel_impl<float, float>(at::TensorIteratorBase &, T2)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | rmsnorm | 0.00496 | 128 | 512 | 2,100,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | reduce | void at::native::reduce_kernel<(int)512, (int)1, at::native::ReduceOp<float, at::native::MeanOps<float, float, float, float>, unsigned int, float, (int)4, (int)4>>(T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 16,384 | rmsnorm | 0.01859 | 512 | 8 | 2,160,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | residual | 0.00342 | 128 | 1 | 5,890 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, at::native::rsqrt_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | rmsnorm | 0.00365 | 128 | 1 | 7,170 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | mul | 0.00608 | 128 | 2,048 | 2,110,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | cast | void at::native::vectorized_elementwise_kernel<(int)8, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | cast | 0.00486 | 128 | 512 | 2,100,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | mul | 0.00659 | 128 | 1,024 | 1,080,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | mul | 0.0065 | 128 | 1,024 | 1,110,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::neg_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 9)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[l... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | neg | 0.00608 | 128 | 512 | 547,330 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | copy | void at::native::<unnamed>::CatArrayBatchedCopy<at::native::<unnamed>::OpaqueType<(unsigned int)2>, unsigned int, (int)4, (int)64, (int)64>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | other | 0.00854 | 512 | 528 | 1,060,000 | 25 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | mul | 0.00637 | 128 | 1,024 | 1,110,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<c10::BFloat16>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | residual | 0.00666 | 128 | 1,024 | 2,120,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 131,072 | mul | 0.0063 | 128 | 256 | 321,790 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::neg_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 9)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[l... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 65,536 | neg | 0.00624 | 128 | 128 | 154,110 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | copy | void at::native::<unnamed>::CatArrayBatchedCopy<at::native::<unnamed>::OpaqueType<(unsigned int)2>, unsigned int, (int)4, (int)64, (int)64>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | other | 0.00662 | 512 | 528 | 269,310 | 25 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 131,072 | mul | 0.00614 | 128 | 256 | 322,050 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<c10::BFloat16>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 131,072 | residual | 0.00627 | 128 | 256 | 551,170 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | copy | void at::native::<unnamed>::CatArrayBatchedCopy<at::native::<unnamed>::OpaqueType<(unsigned int)2>, unsigned int, (int)4, (int)64, (int)64>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | other | 0.00691 | 512 | 528 | 269,310 | 25 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | copy | void at::native::<unnamed>::CatArrayBatchedCopy<at::native::<unnamed>::OpaqueType<(unsigned int)2>, unsigned int, (int)4, (int)64, (int)64>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | other | 0.0065 | 512 | 528 | 269,310 | 25 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::AUnaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | mul | 0.00387 | 128 | 512 | 1,050,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::CUDAFunctor_add<c10::BFloat16>, std::array<char *, (unsigned long)3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | residual | 0.00502 | 128 | 512 | 2,100,000 | 34 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | cast | void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.00608 | 128 | 1,024 | 1,080,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, void at::native::<unnamed>::pow_tensor_scalar_kernel_impl<float, float>(at::TensorIteratorBase &, T2)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | rmsnorm | 0.00483 | 128 | 512 | 2,100,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | reduce | void at::native::reduce_kernel<(int)512, (int)1, at::native::ReduceOp<float, at::native::MeanOps<float, float, float, float>, unsigned int, float, (int)4, (int)4>>(T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 16,384 | rmsnorm | 0.01837 | 512 | 8 | 2,160,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | residual | 0.00355 | 128 | 1 | 5,890 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, at::native::rsqrt_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | rmsnorm | 0.00358 | 128 | 1 | 6,910 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | mul | 0.00605 | 128 | 2,048 | 2,110,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | cast | void at::native::vectorized_elementwise_kernel<(int)8, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | cast | 0.00394 | 128 | 512 | 2,100,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | mul | 0.0063 | 128 | 1,024 | 1,080,000 | 18 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::<unnamed>::silu_kernel(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 6)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | silu | 0.00608 | 128 | 1,600 | 3,290,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | mul | 0.00589 | 128 | 1,600 | 6,560,000 | 40 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::AUnaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | mul | 0.004 | 128 | 512 | 1,050,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)8, at::native::CUDAFunctor_add<c10::BFloat16>, std::array<char *, (unsigned long)3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | residual | 0.00509 | 128 | 512 | 2,100,000 | 34 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | cast | void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.00634 | 128 | 1,024 | 1,080,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, void at::native::<unnamed>::pow_tensor_scalar_kernel_impl<float, float>(at::TensorIteratorBase &, T2)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | rmsnorm | 0.00483 | 128 | 512 | 2,100,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | reduce | void at::native::reduce_kernel<(int)512, (int)1, at::native::ReduceOp<float, at::native::MeanOps<float, float, float, float>, unsigned int, float, (int)4, (int)4>>(T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 16,384 | rmsnorm | 0.0184 | 512 | 8 | 2,160,000 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | residual | 0.00368 | 128 | 1 | 5,890 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::vectorized_elementwise_kernel<(int)4, at::native::rsqrt_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | rmsnorm | 0.00352 | 128 | 1 | 7,170 | 32 |
granite-3.0-8b-instruct_prefill | H100 | Granite-8B | elementwise | void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | mul | 0.00582 | 128 | 2,048 | 2,140,000 | 18 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.