00:00 <miek> also different

00:02 <FFY00> thanks :)

08:00 <tnt> FFY00: https://github.com/smunaut/ecp5-speed/blob/master/results.txt

08:00 <tpb> Title: ecp5-speed/results.txt at master · smunaut/ecp5-speed · GitHub (at github.com)

08:01 <tnt> FFY00: So far observed speed is more related to measured Vcore than the speed on the package, but the sample size is rather small.

12:33 <FFY00> thank you!

12:33 <FFY00> really helpful

17:03 <HackerFoo> Bitstream generation is nearly 4x faster with node reordering. I don't know why.

17:03 <HackerFoo> With node reordering: https://hydra.vtr.tools/build/11277/download/1/meta.json

17:03 <HackerFoo> Without node reordering: https://hydra.vtr.tools/build/11214/download/1/meta.json

17:07 <HackerFoo> I see the same thing for picosoc as well, 12 seconds with, 48 seconds without

17:17 <HackerFoo> I think I might know why

17:23 <andrewb1999> litghost/acomodi: In some partition region tests I am running into the error "Message: No more sub tiles available for initial placement.". Do you think this is related to heterogeneous tiles or something else?

19:28 <andrewb1999> Looks like the sub tile issue has to do with BRAMs

19:28 <andrewb1999> Has anyone encountered this before?

20:32 <andrewb1999> Is there a reason BRAM_R tiles aren't added to architectures?