johnnyffm / poc1to2converter Goto Github PK

During my conversions, on the same drives, plots of differing sizes take different amounts of time to complete, to the tune of 8 hrs vs 23 hrs. This screenshot shows 16 vs 21.

Self-indictment 😁 current version (v2.3) has a bug when trying to convert files >8.35TB with memory limit set to 4096MB.

The issue is not critical, your plot file is 100% ok.

I‘ll not be able to fix this until weekend, however, there is two ways to workaround:

(1) reduce your memory limit a couple of MB. 4090 should do the trick.

(2) use v.2.21

Best,

Johnny

Do we have to convert the plots from PoC1 to PoC2? any benefit from this conversion? Do we received less reward with old plots file?

I have 513 TB of plot files, most of which are 14.5 TB plots made by striping 2 8TB disks on my LSI Raid controllers. When I try to run Poc1Poc2Conv.exe on these files, a second after I click Start Conversion, the program just exits. I have a 1 7.27 TB plot that Poc1Poc2Conv.exe seems to be working fine on.

Is there problem supporting > 8 TB plot files?

Running v.2.3
Disk to disk conversion of file XXXXXXXXXXXX_0_2661664_2661664
I/O Error -Seek to read
Scoop: 4095 The parameter is incorrect
This happens as soon as it starts Processing Scoop Pairs 1/2048

Would be nice to add command line arguments to start the progress from a batch file.

Was trying to convert while mining and the converter crashed when it tried to open a file that was locked by the miner. You can repro it when mining and converting smaller file at the same time (< 500MB). Propose to just skip the file in use.

I have 108 plot files totaling 418 TB. I'm running the latest version of your converter with great success, but a 2TB plot takes about 11 hours to convert inline. So at this rate, it will take me 95 days to convert all my plots.

Will a parallel version be released that will allow converting a few dozen plots at once?

My server has dual E5-2680 V4 CPUs and 64GB of RAM, and all 108 drives are directly connected via SAS2 expanders via 4 x 4 x 6Gbps channel HBAs, so I have the I/O to convert 20-30 plots at once.

This converter and forked Blago miner are otherwise working great. Thanks for all the work putting this together!

Windows 7 x64, Ryzen 7 1700 CPU, 16 GB RAM. There is nothing going on in converter's window when when the "Speed boost" checkbox is checked. The process seems to be stucked at 0%. I opened second instance of converter to convert plot files on another disk without the "speed boost" checkbox, and it is simply working as it should to.

This error keeps coming up on several plot sets (different drives/controllers), all of which have been working flawlessly for a year or more. Can't believe its HW related.

IO Error Seek to Read
Scoop: 4095 The parameter is incorrect

It seems to come up on the second or later plot on a drive and fails the convert session, all remaining plots on the drive come up with the same error if tried again. The plots are not sequential on the individual drives - if that matters. Only 1 out of 5 drives attempted has completed so far....not good. Because the error fails all the other plots on the drive, I have no choice but to re-plot those from scratch.

I'm running Win10_1709, 32GB mem, i5, multiple single disks, running 2 instances of your converter and mining with latest JMiner (GPU based). Performance monitor shows no bottlenecks other than the drives themselves. Hardly using 5GB of RAM when running. Drives being converted were removed from active mining path list.

Any ideas?

I have converted my plots files with Poc1to2Converter v.2.3 on 2 different PCs. blagominer-v1.170911_AVX says about "REALLY BAD deadlines" on both these PCs. Is this something that I should worry about?

Judging by the numbers of nonces the problem is related to different plot files on different disks.

My miner unfortunately suffered from a power outage last night, corrupting 6 TB of plots while converting POC1 -> 2 was currently in progress midway.
To my understanding how conversion works, wouldn't it be sufficient to mark the current scoop (or block of scoops) as nCurrent somewhere on any disk (or the disk in progress to get converted) so whenever conversion fails, it can simply start conversion from the next scoop/block of scoops? With that, one can continue the conversion process and only two scoops or block of scoops are corrupted and not "almost a whole plotfile".