This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TMS320C6678: VLIB performance

Part Number: TMS320C6678

Hi,

My customer is now interested in VLIB function (VLIB_neuralNet_tile). And they would like to know what is meant by "cyc/pt" in the VLIB_c66x_TestReport.html. Do you have any suggestion ?

 ---------------------------------------------------------------------------------------------------------------------------------------------
                   VLIB_neuralNet_tile testing starts.
    ---------------------------------------------------------------------------------------------------------------------------------------------
    | No | Status | Num pt  | Opt Compute  | NatC Compute    |I/O     |  Opt Comp || IO  | Description
    |    |        |         |  cyc/pt      |  cyc/pt         | cyc/pt |  Total cyc/pt    |            
    ---------------------------------------------------------------------------------------------------------------------------------------------
    |  0 | PASS   |     512 |       418.06 |          347.62 |   0.00 |         418.06   | Sequentially generated input | Opt results compared to NatC
    |    |        |         |              |                 |        |                  | results | inCols=32, inRows=1, outCols=16, outRows=1,
    |    |        |         |              |                 |        |                  | numCh=32
    |  1 | PASS   |    1024 |       399.06 |          360.62 |   0.00 |         399.06   | Sequentially generated input | Opt results compared to NatC
    |    |        |         |              |                 |        |                  | results | inCols=64, inRows=1, outCols=16, outRows=1,
    |    |        |         |              |                 |        |                  | numCh=64
    |  2 | PASS   |     512 |       411.54 |          341.09 |   0.00 |         411.54   | Sequentially generated input | Opt results compared to NatC
    |    |        |         |              |                 |        |                  | results | inCols=32, inRows=1, outCols=16, outRows=1,
    |    |        |         |              |                 |        |                  | numCh=32
    |  3 | PASS   |    2048 |       397.53 |          359.31 |   0.00 |         397.53   | Sequentially generated input | Opt results compared to NatC
    |    |        |         |              |                 |        |                  | results | inCols=64, inRows=1, outCols=32, outRows=1,
    |    |        |         |              |                 |        |                  | numCh=64
    |  4 | PASS   |      92 |       726.10 |         1503.99 |   0.00 |         726.10   | Sequentially generated input | Opt results compared to NatC
    |    |        |         |              |                 |        |                  | results | inCols=1152, inRows=1, outCols=92, outRows=1,
    |    |        |         |              |                 |        |                  | numCh=1
    ---------------------------------------------------------------------------------------------------------------------------------------------



Best Regards,

Naoki Kawada

VLIB_neuralNet_tile