Maximum size for persisting L2 cache

Accelerated Computing CUDA CUDA Programming and Performance
1

I am trying to use persistent L2 cache. Before setting cache pollacy, I checked device property to view possible size of persistent cache. However, it looks wired that prints 0 byte as an result. Is there something wrong here? Or, my device does not support persistent cache?

Here is sample code and result

		// Read GPU properties
		cudaDeviceProp properties;
    	int device_idx;
    	cudaError_t result = cudaGetDevice(&device_idx);

		if (result != cudaSuccess) {
		throw std::runtime_error("cudaGetDevice() API call failed.");
		}

    	result = cudaGetDeviceProperties(&properties, device_idx);

		tlog::info() << "GPU: " << properties.name << "[" << device_idx << "]";
		tlog::info() << "L2 Cache Size: " << properties.l2CacheSize / 1024 / 1024 << " MB";
		tlog::info() << "Max Persistent L2 Cache Size: " << properties.persistingL2CacheMaxSize / 1024 / 1024 << " MB";

14:23:37 INFO     GPU: Quadro RTX 5000[0]
14:23:37 INFO     L2 Cache Size: 4 MB
14:23:37 INFO     Max Persistent L2 Cache Size: 0 MB

Here is device information using deviceQuery (demo program)

$ /usr/local/cuda-11.7/extras/demo_suite/deviceQuery 
/usr/local/cuda-11.7/extras/demo_suite/deviceQuery Starting...

 CUDA Device Query (Runtime API) version (CUDART static linking)

Detected 3 CUDA Capable device(s)

Device 0: "Quadro RTX 5000"
  CUDA Driver Version / Runtime Version          11.4 / 11.7
  CUDA Capability Major/Minor version number:    7.5
  Total amount of global memory:                 16125 MBytes (16908615680 bytes)
  (48) Multiprocessors, ( 64) CUDA Cores/MP:     3072 CUDA Cores
  GPU Max Clock rate:                            1815 MHz (1.81 GHz)
  Memory Clock rate:                             7001 Mhz
  Memory Bus Width:                              256-bit
  L2 Cache Size:                                 4194304 bytes
  Maximum Texture Dimension Size (x,y,z)         1D=(131072), 2D=(131072, 65536), 3D=(16384, 16384, 16384)
  Maximum Layered 1D Texture Size, (num) layers  1D=(32768), 2048 layers
  Maximum Layered 2D Texture Size, (num) layers  2D=(32768, 32768), 2048 layers
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total number of registers available per block: 65536
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  1024
  Maximum number of threads per block:           1024
  Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
  Max dimension size of a grid size    (x,y,z): (2147483647, 65535, 65535)
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and kernel execution:          Yes with 3 copy engine(s)
  Run time limit on kernels:                     No
  Integrated GPU sharing Host Memory:            No
  Support host page-locked memory mapping:       Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  Device supports Unified Addressing (UVA):      Yes
  Device supports Compute Preemption:            Yes
  Supports Cooperative Kernel Launch:            Yes
  Supports MultiDevice Co-op Kernel Launch:      Yes
  Device PCI Domain ID / Bus ID / location ID:   0 / 23 / 0
  Compute Mode:
     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

Device 1: "Quadro RTX 5000"
  CUDA Driver Version / Runtime Version          11.4 / 11.7
  CUDA Capability Major/Minor version number:    7.5
  Total amount of global memory:                 16117 MBytes (16900227072 bytes)
  (48) Multiprocessors, ( 64) CUDA Cores/MP:     3072 CUDA Cores
  GPU Max Clock rate:                            1815 MHz (1.81 GHz)
  Memory Clock rate:                             7001 Mhz
  Memory Bus Width:                              256-bit
  L2 Cache Size:                                 4194304 bytes
  Maximum Texture Dimension Size (x,y,z)         1D=(131072), 2D=(131072, 65536), 3D=(16384, 16384, 16384)
  Maximum Layered 1D Texture Size, (num) layers  1D=(32768), 2048 layers
  Maximum Layered 2D Texture Size, (num) layers  2D=(32768, 32768), 2048 layers
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total number of registers available per block: 65536
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  1024
  Maximum number of threads per block:           1024
  Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
  Max dimension size of a grid size    (x,y,z): (2147483647, 65535, 65535)
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and kernel execution:          Yes with 3 copy engine(s)
  Run time limit on kernels:                     No
  Integrated GPU sharing Host Memory:            No
  Support host page-locked memory mapping:       Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  Device supports Unified Addressing (UVA):      Yes
  Device supports Compute Preemption:            Yes
  Supports Cooperative Kernel Launch:            Yes
  Supports MultiDevice Co-op Kernel Launch:      Yes
  Device PCI Domain ID / Bus ID / location ID:   0 / 115 / 0
  Compute Mode:
     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

Device 2: "Quadro RTX 5000"
  CUDA Driver Version / Runtime Version          11.4 / 11.7
  CUDA Capability Major/Minor version number:    7.5
  Total amount of global memory:                 16125 MBytes (16908615680 bytes)
  (48) Multiprocessors, ( 64) CUDA Cores/MP:     3072 CUDA Cores
  GPU Max Clock rate:                            1815 MHz (1.81 GHz)
  Memory Clock rate:                             7001 Mhz
  Memory Bus Width:                              256-bit
  L2 Cache Size:                                 4194304 bytes
  Maximum Texture Dimension Size (x,y,z)         1D=(131072), 2D=(131072, 65536), 3D=(16384, 16384, 16384)
  Maximum Layered 1D Texture Size, (num) layers  1D=(32768), 2048 layers
  Maximum Layered 2D Texture Size, (num) layers  2D=(32768, 32768), 2048 layers
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total number of registers available per block: 65536
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  1024
  Maximum number of threads per block:           1024
  Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
  Max dimension size of a grid size    (x,y,z): (2147483647, 65535, 65535)
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and kernel execution:          Yes with 3 copy engine(s)
  Run time limit on kernels:                     No
  Integrated GPU sharing Host Memory:            No
  Support host page-locked memory mapping:       Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  Device supports Unified Addressing (UVA):      Yes
  Device supports Compute Preemption:            Yes
  Supports Cooperative Kernel Launch:            Yes
  Supports MultiDevice Co-op Kernel Launch:      Yes
  Device PCI Domain ID / Bus ID / location ID:   0 / 166 / 0
  Compute Mode:
     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >
> Peer access from Quadro RTX 5000 (GPU0) -> Quadro RTX 5000 (GPU1) : Yes
> Peer access from Quadro RTX 5000 (GPU0) -> Quadro RTX 5000 (GPU2) : Yes
> Peer access from Quadro RTX 5000 (GPU1) -> Quadro RTX 5000 (GPU0) : Yes
> Peer access from Quadro RTX 5000 (GPU1) -> Quadro RTX 5000 (GPU2) : Yes
> Peer access from Quadro RTX 5000 (GPU2) -> Quadro RTX 5000 (GPU0) : Yes
> Peer access from Quadro RTX 5000 (GPU2) -> Quadro RTX 5000 (GPU1) : Yes

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 11.4, CUDA Runtime Version = 11.7, NumDevs = 3, Device0 = Quadro RTX 5000, Device1 = Quadro RTX 5000, Device2 = Quadro RTX 5000
Result = PASS
2

Persistent L2 became available from Cuda Capability 8.0.

1 Like
3

How can I upgrade it over 8.0?

4

Buy an Ampere, Ada Lovelace or Hopper based card - it’s a hardware thing.

5

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.