Hello there Joe! I've been looking everywhere trying to find where exactly I can modify the MPU PLL. If you could please point us in the right direction that would be great. Thank you.
On Thursday, March 1, 2012 at 3:55:15 AM UTC-8, Uncle Joe wrote: > > Fellow high performance computing enthusiasts, > > I just overclocked my dog from 720MHz to 1152 MHz (only MPU PLL) ! > Take that you lazy dog! Get those ARMs moving! Crunch those numbers > like you mean it. > > Strangely, increasing the TPS65217's DCDC3 voltage from 1.1V to 1.3V > didn't add more overclocking room. I just realized OMAP3 is an > overclocker's dream. So many PLLs and clock dividers to play with. > I've also doubled the clock to the SGX graphics unit to 400MHz without > problem. Too bad I don't have the HDMI cape so that I can make a video > showing how much faster the graphics go. Next, I will attempt to > overclock the DDR2 memory from 266 MhZ to its 400MHz spec. > > Has anyone pushed their board beyond 1.2 GHz? Anyways, I'd better > hide. Gerald Coley is going to come after me with a shotgun for > abusing his pet like he did to that cheating PCB vendor. Yikes! > > > -----------------------before - 720 MHz------------------------------ > > LINPACK benchmark, Single precision. > Machine precision: 6 digits. > Array size 200 X 200. > Average rolled and unrolled performance: > > Reps Time(s) DGEFA DGESL OVERHEAD KFLOPS > ---------------------------------------------------- > 16 0.79 87.34% 5.06% 7.59% 30100.461 > 32 1.60 88.13% 3.75% 8.12% 29895.688 > 64 3.18 88.68% 2.52% 8.81% 30308.049 > 128 6.37 88.70% 2.51% 8.79% 30255.896 > 256 12.73 88.53% 2.59% 8.88% 30308.035 > 512 25.47 89.05% 2.20% 8.76% 30255.877 > 1024 50.93 88.65% 2.61% 8.74% 30255.879 > > --------------------------after - 1152 > MHz-------------------------------- > > LINPACK benchmark, Single precision. > Machine precision: 6 digits. > Array size 200 X 200. > Average rolled and unrolled performance: > > Reps Time(s) DGEFA DGESL OVERHEAD KFLOPS > ---------------------------------------------------- > 32 0.99 94.95% 1.01% 4.04% 46259.645 > 64 1.96 90.31% 2.04% 7.65% 48559.863 > 128 3.94 90.10% 3.30% 6.60% 47768.129 > 256 7.86 89.69% 2.93% 7.38% 48293.012 > 512 15.73 89.96% 2.48% 7.57% 48359.477 > 1024 31.45 89.79% 2.80% 7.41% 48293.090 > 2048 62.89 89.89% 2.61% 7.51% 48351.172 > > > (synthetic test I made, measuring only raw arithmetic throughput) > > --------------------------before - 720 > MHz------------------------------- > 1.15017G int adds/s > 0.359117G int multiplies/s > 0.102649G float adds/s > 2.87421G neon int32 adds/s > 0.718913G neon int32 multiplies/s > 1.43731G neon float adds/s > 1.43813G neon float multiplies/s > 1.43747G neon float madds/s > > --------------------------after - 1152 > MhZ----------------------------------- > 1.84078G int adds/s > 0.574921G int multiplies/s > 0.164397G float adds/s > 4.60351G neon int32 adds/s > 1.15064G neon int32 multiplies/s > 2.30213G neon float adds/s > 2.2995G neon float multiplies/s > 2.30176G neon float madds/s > > -- For more options, visit http://beagleboard.org/discuss --- You received this message because you are subscribed to the Google Groups "BeagleBoard" group. To unsubscribe from this group and stop receiving emails from it, send an email to beagleboard+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
