tree 3acadb88634106631bf4494bbdc2c45d9b69fcc2
parent de6481ccde749b33a62731d380aa0dc1292b3a2d
author chiyotsai <chiyotsai@google.com> 1661381527 -0700
committer Chi Yo Tsai <chiyotsai@google.com> 1661564347 +0000

RTC: Add SSE4_1 optimization for aom_vector_var

This commit also fixes a potential integer overflow issue in the C code
when the function is applied to BLOCK_128X128.

Performance:
Functional Level:
| Width | C Code | SSE4_1 | Gain |
|-------|--------|--------|------|
|   16  |  42430 |  25017 | 1.70 |
|   32  |  69805 |  31255 | 2.28 |
|   64  | 126035 |  43659 | 2.80 |
|  128  | 238695 |  72269 | 3.30 |

Encoder Level:
| SPD_SET | TESTSET  | AVG_PSNR | OVR_PSNR |  SSIM   | ENC_T |
|---------|----------|----------|----------|---------|-------|
|    7    |   rtc    | +0.000%  | +0.000%  | +0.000% | -1.2% |
|    7    | rtc_derf | +0.000%  | +0.000%  | +0.000% | -1.2% |
|---------|----------|----------|----------|---------|-------|
|    8    |   rtc    | +0.000%  | +0.000%  | +0.000% | -1.2% |
|    8    | rtc_derf | +0.000%  | +0.000%  | +0.000% | -0.8% |
|---------|----------|----------|----------|---------|-------|
|    9    |   rtc    | +0.000%  | +0.000%  | +0.000% | -0.1% |
|    9    | rtc_derf | +0.000%  | +0.000%  | +0.000% | -0.1% |
|---------|----------|----------|----------|---------|-------|
|   10    |   rtc    | +0.000%  | +0.000%  | +0.000% | -0.1% |
|   10    | rtc_derf | +0.000%  | +0.000%  | +0.000% | -0.1% |

Change-Id: If02b50a037b9f4b87fc57d471a6998130f575d41