| <<<Back 1 day (to 2020/10/28) | Fwd 1 day (to 2020/10/30) >>> | 20201029 |
paulgardiner | Experimented with high-security redaction again. Same test page. With the floating-point test data, the page took 47s; neon made no difference. With the integerised test data, 32s; neon cut it to 28s. | 15:47.33 |
| Results still seem to be accurate. | 15:48.00 |
kens | paulgardiner: are you intending to be in #ghostscript ? | 15:48.21 |
paulgardiner | No, not really. :-) Still it looks like it will have caught the intended people. | 15:49.43 |
kens | If you're after Robin_Watts_ you should probably ping him :-) | 15:50.17 |
paulgardiner | Good point. | 15:51.14 |
Robin_Watts_ | Hi paul, sorry, didn't see you in this group. | 15:51.49 |
paulgardiner | Robin_Watts_: see comment on redactions a few lines up | 15:51.50 |
| No probs. Wasn't looking for a response. | 15:52.14 |
Robin_Watts_ | paulgardiner: So a 12% saving or so? sounds plausible. | 15:54.09 |
paulgardiner | Yeah. I guess it's not certain I'm getting all neon might offer. I had to remove the -mneon compiler flag and retain just -DHAVE_NEON to compile on iOS | 15:55.59 |
| for iOS | 15:56.03 |
| <<<Back 1 day (to 2020/10/28) | Forward 1 day (to 2020/10/30)>>> | |