Benchmark Result¶
aes256CbcDecrypt throughput¶
To profile performance of aes256CbcDecrypt, we prepare a datapack of 32K messages, each message is 1Kbyte. We have 1 kernels, each kernel has 4 PUs. Kernel utilization and throughput is shown in table below.
| Frequency | LUT | REG | BRAM | URAM | DSP | Throughput |
|---|---|---|---|---|---|---|
| 286MHz | 203,595(11.8%) | 312,900(9.1%) | 761(28.3%) | 0 | 29(0.3%) | 4.7GB/s |
aes256CbcEncrypt throughput¶
To profile performance of aes256CbcEncrypt, we prepare a datapack of 96K messages, each message is 1Kbyte. We have 4 kernels, each kernel has 12 PUs. Kernel utilization and throughput is shown in table below.
| Frequency | LUT | REG | BRAM | URAM | DSP | Throughput |
|---|---|---|---|---|---|---|
| 224MHz | 1,059,093(61.3%) | 1,010,145(34.9%) | 654(24.3%) | 0 | 152(1.3%) | 5.5GB/s |
rc4 throughput¶
To profile performance of rc4, we prepare a datapack of 24 messages, each message is 2Mbyte. We have 4 kernels, each kernel has 12 PUs. Kernel utilization and throughput is shown in table below.
| Frequency | LUT | REG | BRAM | URAM | DSP | Throughput |
|---|---|---|---|---|---|---|
| 147MHz | 1,126,259(73.0%) | 1,120,505(34.9%) | 640 (44.0%) | 0 | 216(1.8%) | 3.0GB/s |
hmacSha1 throughput¶
To profile performance of hmacSha1, we prepare a datapack of 512K messages, each message is 1Kbyte, key length is 256bits. We have 4 kernels, each kernel has 16 PUs. Kernel utilization and throughput is shown in table below.
| Frequency | LUT | REG | BRAM | URAM | DSP | Throughput |
|---|---|---|---|---|---|---|
| 227 MHz | 959,078(55.5%) | 1,794,522(52.0%) | 777 (28.9%) | 0 | 56(0.5%) | 8.0 GB/s |
crc32 throughput¶
To profile performance of crc32, we prepare a datapack of 268,435,456 byte messages as kernel input. Base on U50, We have 1 kernel, each kernel has 1 PU. Kernel utilization and throughput is shown in table below.
| Frequency | LUT | REG | BRAM | URAM | DSP | Throughput |
|---|---|---|---|---|---|---|
| 300 MHz | 5,322(0.69%) | 10,547(0.65%) | 16 (1.37%) | 0 | 0 | 4.7 GB/s |
adler32 throughput¶
To profile performance of adler32, we prepare a datapack of 268,435,456 byte messages as kernel input. Base on U50, We have 1 kernel, each kernel has 1 PU. Kernel utilization and throughput is shown in table below.
| Frequency | LUT | REG | BRAM | URAM | DSP | Throughput |
|---|---|---|---|---|---|---|
| 262 MHz | 6,348(0.83%) | 12,232(0.76%) | 16 (1.37%) | 0 | 0 | 4.1 GB/s |