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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::CUDAFunctor_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | residual | 0.00634 | 128 | 1,024 | 2,110,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::exp_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 4,096 | other | 0.00314 | 128 | 8 | 18,300 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::BinaryFunctor<float, float, float, 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 | 0.00554 | 128 | 1,024 | 1,070,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::CUDAFunctor_add<float>, std::array<char *, 3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 131,072 | residual | 0.00605 | 128 | 256 | 2,100,000 | 40 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.00528 | 128 | 1,024 | 1,060,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::exp_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | other | 0.00272 | 128 | 1 | 12,160 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::BinaryFunctor<float, float, float, 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 | 0.00762 | 128 | 2,048 | 2,110,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::CUDAFunctor_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 4,096 | residual | 0.0031 | 128 | 8 | 18,720 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::exp_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,024 | other | 0.00234 | 128 | 2 | 15,520 | 30 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::BinaryFunctor<float, float, float, 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 | 0.00534 | 128 | 1,024 | 1,070,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::CUDAFunctor_add<float>, std::array<char *, 3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | residual | 0.0097 | 128 | 512 | 4,210,000 | 40 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | cast | 0.00794 | 128 | 2,048 | 2,110,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::vectorized_elementwise_kernel<4, at::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | cast | 0.00627 | 128 | 512 | 2,100,000 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::FillFunctor<c10::BFloat16>, std::array<char *, 1>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | fill | 0.00256 | 128 | 512 | 3,740 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::vectorized_elementwise_kernel<4, at::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | cast | 0.00627 | 128 | 512 | 2,110,000 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::unrolled_elementwise_kernel<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>, 4, TrivialOffsetCalculator<1, unsigned int>, TrivialOffsetCalculator<1, unsigned in... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.00826 | 128 | 1,024 | 1,080,000 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, void at::<unnamed>::pow_tensor_scalar_kernel_impl<float, float>(at::TensorIteratorBase &, T2)::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | rmsnorm | 0.00608 | 128 | 512 | 2,100,000 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | reduce | void at::reduce_kernel<512, 1, at::ReduceOp<float, at::MeanOps<float, float, float, float>, unsigned int, float, 4, 4>>(T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | rmsnorm | 0.0104 | 512 | 256 | 2,140,000 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::CUDAFunctorOnSelf_add<float>, std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 2,048 | residual | 0.00224 | 128 | 4 | 20,860 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::rsqrt_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 2,048 | rmsnorm | 0.00237 | 128 | 4 | 22,430 | 24 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::BinaryFunctor<float, float, float, 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 | 0.00774 | 128 | 2,048 | 2,120,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::vectorized_elementwise_kernel<4, at::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | cast | 0.00611 | 128 | 512 | 2,100,000 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 4, void at::gpu_kernel_impl_nocast<at::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, 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 | 0.00765 | 128 | 1,024 | 1,070,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::unrolled_elementwise_kernel<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>, 4, TrivialOffsetCalculator<1, unsigned int>, TrivialOffsetCalculator<1, unsigned in... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.00829 | 128 | 1,024 | 1,080,000 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::<unnamed>::silu_kernel(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | silu | 0.00694 | 128 | 512 | 2,110,000 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::unrolled_elementwise_kernel<at::BinaryFunctor<float, float, float, binary_internal::MulFunctor<float>>, std::array<char *, 3>, 4, TrivialOffsetCalculator<2, unsigned int>, TrivialOffsetCalculator<1, unsigned int>, memory::LoadWithCast<2>, memory::StoreWithCast<1>>(int, T1, T2, T4, T5, T6, T7) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | mul | 0.01242 | 128 | 1,024 | 3,180,000 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::vectorized_elementwise_kernel<4, at::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | cast | 0.00621 | 128 | 512 | 2,100,000 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::CUDAFunctor_add<c10::BFloat16>, std::array<char *, 3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | residual | 0.0063 | 128 | 512 | 2,110,000 | 40 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::unrolled_elementwise_kernel<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>, 4, TrivialOffsetCalculator<1, unsigned int>, TrivialOffsetCalculator<1, unsigned in... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.00835 | 128 | 1,024 | 1,080,000 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, void at::<unnamed>::pow_tensor_scalar_kernel_impl<float, float>(at::TensorIteratorBase &, T2)::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | rmsnorm | 0.00605 | 128 | 512 | 2,100,000 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | reduce | void at::reduce_kernel<512, 1, at::ReduceOp<float, at::MeanOps<float, float, float, float>, unsigned int, float, 4, 4>>(T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 16,384 | rmsnorm | 0.0167 | 512 | 8 | 2,140,000 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::CUDAFunctorOnSelf_add<float>, std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | residual | 0.00214 | 128 | 1 | 4,990 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::rsqrt_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | rmsnorm | 0.0023 | 128 | 1 | 6,560 | 24 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::BinaryFunctor<float, float, float, 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 | 0.00752 | 128 | 2,048 | 2,110,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::unrolled_elementwise_kernel<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>, 4, TrivialOffsetCalculator<1, unsigned int>, TrivialOffsetCalculator<1, unsigned in... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 4,096 | cast | 0.00518 | 128 | 8 | 40,830 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::CUDAFunctorOnSelf_add<float>, std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 2,048 | residual | 0.00221 | 128 | 4 | 20,860 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::BinaryFunctor<float, float, float, 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 | 0.00768 | 128 | 2,048 | 2,120,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::vectorized_elementwise_kernel<4, at::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | cast | 0.00627 | 128 | 512 | 2,100,000 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::<unnamed>::silu_kernel(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 6)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | silu | 0.00915 | 128 | 1,536 | 3,430,000 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, binary_internal::MulFunctor<float>>, std::array<char *, 3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | mul | 0.01443 | 128 | 1,536 | 6,840,000 | 40 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::CUDAFunctor_add<c10::BFloat16>, std::array<char *, 3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | residual | 0.00621 | 128 | 512 | 2,110,000 | 40 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::unrolled_elementwise_kernel<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>, 4, TrivialOffsetCalculator<1, unsigned int>, TrivialOffsetCalculator<1, unsigned in... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.00835 | 128 | 1,024 | 1,080,000 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, void at::<unnamed>::pow_tensor_scalar_kernel_impl<float, float>(at::TensorIteratorBase &, T2)::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | rmsnorm | 0.00611 | 128 | 512 | 2,100,000 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | reduce | void at::reduce_kernel<512, 1, at::ReduceOp<float, at::MeanOps<float, float, float, float>, unsigned int, float, 4, 4>>(T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 16,384 | rmsnorm | 0.01674 | 512 | 8 | 2,140,000 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::CUDAFunctorOnSelf_add<float>, std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | residual | 0.00214 | 128 | 1 | 4,990 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::rsqrt_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | rmsnorm | 0.0023 | 128 | 1 | 6,560 | 24 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::BinaryFunctor<float, float, float, 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 | 0.00762 | 128 | 2,048 | 2,110,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::unrolled_elementwise_kernel<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>, 4, TrivialOffsetCalculator<1, unsigned int>, TrivialOffsetCalculator<1, unsigned in... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 4,096 | cast | 0.00522 | 128 | 8 | 40,830 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::CUDAFunctorOnSelf_add<float>, std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 2,048 | residual | 0.00221 | 128 | 4 | 20,860 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::BinaryFunctor<float, float, float, 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 | 0.00765 | 128 | 2,048 | 2,120,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::vectorized_elementwise_kernel<4, at::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | cast | 0.00608 | 128 | 512 | 2,100,000 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::FillFunctor<c10::BFloat16>, std::array<char *, 1>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 16,384 | fill | 0.00186 | 128 | 32 | 3,740 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::elementwise_kernel<128, 4, void at::gpu_kernel_impl_nocast<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 12)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | cast | 0.02963 | 128 | 2,048 | 2,110,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 4, void at::gpu_kernel_impl_nocast<at::<unnamed>::silu_kernel(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 6)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(i... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,048,576 | silu | 0.01302 | 128 | 2,048 | 2,170,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::sigmoid_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 2,048 | other | 0.00307 | 128 | 4 | 22,940 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::unrolled_elementwise_kernel<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>, 4, TrivialOffsetCalculator<1, unsigned int>, TrivialOffsetCalculator<1, unsigned in... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | cast | 0.00282 | 128 | 1 | 32,700 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::exp_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | other | 0.00243 | 128 | 1 | 7,460 | 30 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::neg_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | neg | 0.00214 | 128 | 1 | 4,640 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::unrolled_elementwise_kernel<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>, 4, TrivialOffsetCalculator<1, unsigned int>, TrivialOffsetCalculator<1, unsigned in... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 4,096 | cast | 0.00518 | 128 | 8 | 40,830 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 4, void at::gpu_kernel_impl<at::CUDAFunctor_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 4,096 | residual | 0.00954 | 128 | 8 | 63,330 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::<unnamed>::softplus_kernel(at::TensorIteratorBase &, const c10::Scalar &, const c10::Scalar &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 2,048 | other | 0.00397 | 128 | 4 | 39,300 | 25 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::BinaryFunctor<float, float, float, binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 8,192 | mul | 0.00307 | 128 | 16 | 27,040 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::elementwise_kernel<128, 4, void at::gpu_kernel_impl_nocast<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 12)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.01606 | 128 | 1,024 | 542,080 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::elementwise_kernel<128, 4, void at::gpu_kernel_impl_nocast<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 12)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.01626 | 128 | 1,024 | 541,630 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, binary_internal::MulFunctor<float>>, std::array<char *, 3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | mul | 0.00448 | 128 | 512 | 1,060,000 | 40 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | reduce | void at::reduce_kernel<512, 1, at::ReduceOp<c10::BFloat16, at::func_wrapper_t<c10::BFloat16, at::sum_functor<c10::BFloat16, float, c10::BFloat16>::operator ()(at::TensorIterator &)::[lambda(float, float) (instance 1)]>, unsigned int, c10::BFloat16, 4, 8>>(T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | other | 0.00758 | 512 | 128 | 1,090,000 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::CUDAFunctorOnSelf_add<c10::BFloat16>, std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 2,048 | residual | 0.00227 | 128 | 4 | 15,740 | 30 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::rsqrt_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 3)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 2,048 | rmsnorm | 0.00234 | 128 | 4 | 16,540 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 4, void at::gpu_kernel_impl_nocast<at::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, 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 | 0.00736 | 128 | 1,024 | 1,080,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, binary_internal::MulFunctor<float>>, std::array<char *, 3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | mul | 0.00442 | 128 | 512 | 1,060,000 | 40 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | reduce | void at::reduce_kernel<512, 1, at::ReduceOp<c10::BFloat16, at::func_wrapper_t<c10::BFloat16, at::sum_functor<c10::BFloat16, float, c10::BFloat16>::operator ()(at::TensorIterator &)::[lambda(float, float) (instance 1)]>, unsigned int, c10::BFloat16, 4, 8>>(T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | other | 0.00755 | 512 | 128 | 1,090,000 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::CUDAFunctorOnSelf_add<c10::BFloat16>, std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 2,048 | residual | 0.00224 | 128 | 4 | 15,740 | 30 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::rsqrt_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 3)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 2,048 | rmsnorm | 0.00234 | 128 | 4 | 16,540 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 4, void at::gpu_kernel_impl_nocast<at::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, 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 | 0.00736 | 128 | 1,024 | 1,080,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::elementwise_kernel<128, 4, void at::gpu_kernel_impl_nocast<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 12)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.00758 | 128 | 1,024 | 1,070,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::unrolled_elementwise_kernel<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>, 4, TrivialOffsetCalculator<1, unsigned int>, TrivialOffsetCalculator<1, unsigned in... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.00816 | 128 | 1,024 | 1,080,000 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::elementwise_kernel<128, 4, void at::gpu_kernel_impl_nocast<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 12)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.00752 | 128 | 1,024 | 1,070,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::unrolled_elementwise_kernel<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>, 4, TrivialOffsetCalculator<1, unsigned int>, TrivialOffsetCalculator<1, unsigned in... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.00832 | 128 | 1,024 | 1,080,000 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::elementwise_kernel<128, 4, void at::gpu_kernel_impl_nocast<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 12)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.0161 | 128 | 1,024 | 1,070,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::unrolled_elementwise_kernel<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>, 4, TrivialOffsetCalculator<1, unsigned int>, TrivialOffsetCalculator<1, unsigned in... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | cast | 0.00826 | 128 | 1,024 | 1,080,000 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::elementwise_kernel<128, 4, void at::gpu_kernel_impl_nocast<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 12)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 4,096 | cast | 0.00342 | 128 | 8 | 25,470 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::unrolled_elementwise_kernel<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>, 4, TrivialOffsetCalculator<1, unsigned int>, TrivialOffsetCalculator<1, unsigned in... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 4,096 | cast | 0.00509 | 128 | 8 | 40,830 | 32 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 8,192 | cast | 0.00307 | 128 | 16 | 26,240 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::AUnaryFunctor<float, float, float, binary_internal::MulFunctor<float>>, std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 262,144 | mul | 0.00621 | 128 | 512 | 2,100,000 | 27 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::BinaryFunctor<float, float, float, 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 | 0.00768 | 128 | 2,048 | 2,120,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::BinaryFunctor<float, float, float, 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 | 0.00752 | 128 | 2,048 | 2,120,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::FillFunctor<bool>, std::array<char *, 1>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 1,024 | fill | 0.00182 | 128 | 2 | 4,260 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::CUDAFunctor_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | residual | 0.00464 | 128 | 1,024 | 26,910 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::exp_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 131,072 | other | 0.00442 | 128 | 256 | 1,060,000 | 30 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::BinaryFunctor<float, float, float, binary_internal::MulFunctor<float>>, std::array<char *, 3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 131,072 | mul | 0.00614 | 128 | 256 | 2,100,000 | 42 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::<unnamed>::masked_fill_kernel(at::TensorIterator &, const c10::Scalar &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float, bool) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda... | bf16 | false | null | null | null | 1 | 128 | null | null | null | 524,288 | other | 0.00528 | 128 | 1,024 | 1,060,000 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::neg_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, 2>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 131,072 | neg | 0.00442 | 128 | 256 | 1,050,000 | 29 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | cast | 0.00253 | 128 | 1 | 13,950 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | cast | 0.00262 | 128 | 1 | 13,950 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::BinaryFunctor<float, float, float, binary_internal::MulFunctor<float>>, std::array<char *, 3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | mul | 0.00218 | 128 | 1 | 5,500 | 42 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | reduce | void at::reduce_kernel<128, 4, at::ReduceOp<float, at::func_wrapper_t<float, at::sum_functor<float, float, float>::operator ()(at::TensorIterator &)::[lambda(float, float) (instance 1)]>, unsigned int, float, 4, 4>>(T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 64 | other | 0.00704 | 16 | 1 | 45,980 | 46 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | elementwise | void at::vectorized_elementwise_kernel<4, at::CUDAFunctor_add<float>, std::array<char *, 3>>(int, T2, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | residual | 0.00221 | 128 | 1 | 5,500 | 40 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | cast | 0.00256 | 128 | 1 | 10,110 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | cast | 0.00259 | 128 | 1 | 13,950 | 16 |
Qwen3.5-9B_prefill | RTX2080Ti | Qwen3.5-9B | cast | void at::elementwise_kernel<128, 2, void at::gpu_kernel_impl_nocast<at::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3) | bf16 | false | null | null | null | 1 | 128 | null | null | null | 512 | cast | 0.00323 | 128 | 1 | 18,050 | 16 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.