Performance counter 1 is used for measuring texturing and caching performance as well as FIFO performance.
- Note
GX_PERF1_TC_* can be used to compute the texture cache (TC) miss rate. The TC_CHECK* parameters count how many texture cache lines are accessed for each pixel. In the worst case, for a mipmap, up to 8 cache lines may be accessed to produce one textured pixel. GX_PERF1_TC_MISS counts how many of those accesses missed the texture cache. To compute the miss rate, divide GX_PERF1_TC_MISS by the sum of all four GX_PERF1_TC_CHECK* values.
-
GX_PERF1_VC_* count different vertex cache stall conditions.
◆ GX_PERF1_CALL_REQ
| #define GX_PERF1_CALL_REQ 18 |
Number of lines (32B) read from called display lists.
◆ GX_PERF1_CLOCKS
| #define GX_PERF1_CLOCKS 21 |
Number of GP clocks that have elapsed since the last call to GX_ReadGP1Metric().
◆ GX_PERF1_CP_ALL_REQ
| #define GX_PERF1_CP_ALL_REQ 20 |
Counts all requests (32B/request) from the GP Command Processor (CP). It should be equal to the sum of counts returned by GX_PERF1_FIFO_REQ, GX_PERF1_CALL_REQ, and GX_PERF1_VC_MISS_REQ.
◆ GX_PERF1_FIFO_REQ
| #define GX_PERF1_FIFO_REQ 17 |
Number of lines (32B) read from the GP FIFO.
◆ GX_PERF1_NONE
Disables performance measurement for perf1 and resets the counter.
◆ GX_PERF1_TC_CHECK1_2
| #define GX_PERF1_TC_CHECK1_2 4 |
◆ GX_PERF1_TC_CHECK3_4
| #define GX_PERF1_TC_CHECK3_4 5 |
◆ GX_PERF1_TC_CHECK5_6
| #define GX_PERF1_TC_CHECK5_6 6 |
◆ GX_PERF1_TC_CHECK7_8
| #define GX_PERF1_TC_CHECK7_8 7 |
◆ GX_PERF1_TC_MISS
| #define GX_PERF1_TC_MISS 8 |
Number of texture cache misses in total?
◆ GX_PERF1_TEXELS
| #define GX_PERF1_TEXELS 0 |
Number of texels processed by the GP.
◆ GX_PERF1_TX_IDLE
| #define GX_PERF1_TX_IDLE 1 |
Number of clocks that the texture unit (TX) is idle.
◆ GX_PERF1_TX_MEMSTALL
| #define GX_PERF1_TX_MEMSTALL 3 |
Number of GP clocks the TX unit is stalled waiting for main memory.
◆ GX_PERF1_TX_REGS
| #define GX_PERF1_TX_REGS 2 |
Number of GP clocks spent writing to state registers in the TX unit.
◆ GX_PERF1_VC_ALL_STALLS
| #define GX_PERF1_VC_ALL_STALLS 15 |
◆ GX_PERF1_VC_ELEMQ_FULL
| #define GX_PERF1_VC_ELEMQ_FULL 9 |
◆ GX_PERF1_VC_MEMREQ_FULL
| #define GX_PERF1_VC_MEMREQ_FULL 11 |
◆ GX_PERF1_VC_MISS_REQ
| #define GX_PERF1_VC_MISS_REQ 19 |
Number vertex cache miss request. Each miss requests a 32B transfer from main memory.
◆ GX_PERF1_VC_MISSQ_FULL
| #define GX_PERF1_VC_MISSQ_FULL 10 |
◆ GX_PERF1_VC_MISSREP_FULL
| #define GX_PERF1_VC_MISSREP_FULL 13 |
◆ GX_PERF1_VC_STATUS7
| #define GX_PERF1_VC_STATUS7 12 |
◆ GX_PERF1_VC_STREAMBUF_LOW
| #define GX_PERF1_VC_STREAMBUF_LOW 14 |
◆ GX_PERF1_VERTICES
| #define GX_PERF1_VERTICES 16 |
Number of vertices processed by the GP.