Browse Source

Reinstate more aggressive <2% HW error target for avalon-auto

nfactor-troky
Con Kolivas 12 years ago
parent
commit
3ef0db5658
  1. 2
      ASIC-README
  2. 6
      driver-avalon.c

2
ASIC-README

@ -94,7 +94,7 @@ Avalon commands:
Avalon auto will enable dynamic overclocking gradually increasing and Avalon auto will enable dynamic overclocking gradually increasing and
decreasing the frequency till the highest hashrate that keeps hardware errors decreasing the frequency till the highest hashrate that keeps hardware errors
around 1% is achieved. This WILL run your avalon beyond its normal specification under 2% is achieved. This WILL run your avalon beyond its normal specification
so the usual warnings apply. When avalon-auto is enabled, the avalon-options so the usual warnings apply. When avalon-auto is enabled, the avalon-options
for frequency and timeout are used as the starting point only. for frequency and timeout are used as the starting point only.

6
driver-avalon.c

@ -916,10 +916,10 @@ static void *avalon_send_tasks(void *userdata)
info->auto_nonces <= (AVALON_AUTO_CYCLE * 21 / 20)) { info->auto_nonces <= (AVALON_AUTO_CYCLE * 21 / 20)) {
int total = info->auto_nonces + info->auto_hw; int total = info->auto_nonces + info->auto_hw;
/* Try to keep hw errors ~1% */ /* Try to keep hw errors < 2% */
if (info->auto_hw * 200 < total) if (info->auto_hw * 100 < total)
avalon_inc_freq(info); avalon_inc_freq(info);
else if (info->auto_hw * 100 > total) else if (info->auto_hw * 66 > total)
avalon_dec_freq(info); avalon_dec_freq(info);
} }
avalon_reset_auto(info); avalon_reset_auto(info);

Loading…
Cancel
Save