2018-06-27 12:43:06
this guy is crushing it, must be spending some big $

Shoots
2018-06-27 12:43:10
snow:0quas9p42rwt5vp4j92u9up3h9hjlna8fm5rah4k

Shoots
2018-06-27 15:20:32
When setting the memfield cache size is this only for the memfield? Or is this for the entire miner and all the other stuff it puts into memory?

Shoots
2018-06-27 15:30:41
only for the memfield

Rotonen
2018-06-27 15:35:47
for some reason it keeps stalling when loading into mem

Shoots
2018-06-27 15:35:58
and when I check htop I see this

Shoots
2018-06-27 15:36:59

Shoots
2018-06-27 15:37:17
heapdumpoutofmemory

Shoots
2018-06-27 15:37:33
even though I have xmx set to 18gb

Shoots
2018-06-27 15:48:01
what do you have the precache set to?

Rotonen
2018-06-27 15:48:32
and, as said, get it to start with just one thread at first

Rotonen
2018-06-27 15:48:45
one problem at a time

Rotonen
2018-06-27 15:49:27
precache 1, threads 1

Rotonen
2018-06-27 15:49:37
then keep bumping the precache

Rotonen
2018-06-27 15:49:47
then start bumping the thread count

Rotonen
2018-06-27 16:12:49
same problem

Shoots
2018-06-27 16:12:57
one thread one gb cache

Shoots
2018-06-27 16:13:03

Shoots
2018-06-27 16:15:39
looks like it's taking a long time to load 1gb...

Tyler Boone
2018-06-27 16:17:04
this is what Im using to launch ./bazel-bin/SnowBlossomMiner --jvm_flags=-Xmx185g /var/snowblossom/configs/miner-mainnet.conf

Shoots
2018-06-27 16:18:12
and I keep seeing this outofmemoryerror in my running PID's

Shoots
2018-06-27 16:19:22
I don't understand that last statement

Tyler Boone
2018-06-27 16:20:13
its going now with 1 thread and 1gb loaded into mem, you were right just took a long time

Shoots
2018-06-27 16:23:03
it would be nice if it wasnt trying to load work from the pool while its still loading the memcache

Shoots
2018-06-27 16:23:11
I've noticed on my machine once it gets to about 47gb loaded it hangs for a good minute or two. I haven't looked into it because it doesn't bother me, but I suppose it's some heavy lifting inside the garbage collector

Tyler Boone
2018-06-27 16:23:48
meh... all the threads are blocked waiting, so I don't see a problem

Tyler Boone
2018-06-27 16:25:27
I just mean so you dont see it writing out that its stalled

Shoots
2018-06-27 16:25:51
oh yeah

Tyler Boone
2018-06-27 16:27:19
I actually did want to load it before starting all that nonsense, but because of the design of the FieldScan to support swapping to a higher snowfield without restarting the miner it would have required a lot more piping and I didn't want to fuddle up the code too bad

Tyler Boone
2018-06-27 16:27:58
also the solo miner seems to load faster than the pool miner

Shoots
2018-06-27 16:29:19
!

Tyler Boone
2018-06-27 16:29:25
how many times did you repro that?

Tyler Boone
2018-06-27 16:29:55
well Im just noticing by the approx ten times Ive tried

Shoots
2018-06-27 16:30:10
no clue.

Tyler Boone
2018-06-27 16:30:28
the solo miner will get the first 6gb in before the first stall message, the pool miner takes ten minutes before it gets even a few gb loaded

Shoots
2018-06-27 16:30:49
hard to explain

Shoots
2018-06-27 16:36:19
i’m also +1 on not disk mining before memfield or precache is done - the disk mining saturates io and there is a net loss

Rotonen
2018-06-27 16:37:37
there shouldn't be any disk mining happening before precache

Tyler Boone
2018-06-27 16:37:43
13 min with solo miner now and only 5gb loaded

Shoots
2018-06-27 16:37:52
wtf?

Tyler Boone
2018-06-27 16:38:03
pool miner I mean

Shoots
2018-06-27 16:38:04
51gb loads for me in like 3-4 minutes

Tyler Boone
2018-06-27 16:38:13
its as if the work from the pool is interrupting it

Shoots
2018-06-27 16:39:06
https://github.com/snowblossomcoin/snowblossom/blob/27e9c53e31d7c9588c643f8e7336456cd62e92d8/miner/src/SnowMerkleProof.java#L165 ```
synchronized (this)
```

Tyler Boone
2018-06-27 16:39:18
all the threads should be blocked on that line (except the one loading the precache)

Tyler Boone
2018-06-27 16:39:51
you can do [kill -3 <pid>] to force java to print out the stacktraces

Tyler Boone
2018-06-27 16:40:19
(or use jstack, whichever you prefer)

Tyler Boone
2018-06-27 16:40:51
@Tyler Boone it’s been definitively mining for me before memfield is complete

Rotonen
2018-06-27 16:41:17
@Rotonen, send logs

Tyler Boone
2018-06-27 16:41:36
or does it start mining the early loaded memfield?

Rotonen
2018-06-27 16:41:45
when using memfield (non hybrid) it mines for me before done loading too

Shoots
2018-06-27 16:41:49
no, it waits for the entire precache

Tyler Boone
2018-06-27 16:42:06
rtfc

Fireduck
2018-06-27 16:42:09
when hybrid mining it doesnt have a hashrate until the cache is done loading

Shoots
2018-06-27 16:42:37
@Tyler Boone will do, not near a privileged terminal within the next 6h or so, though

Rotonen
2018-06-27 16:42:46
but its taking a god awful long time to load the cache into mem

Shoots
2018-06-27 16:43:56
@Shoots sure it is the same snowpath between solo and pool?

Fireduck
2018-06-27 16:44:16
maybe the drive is sleepy

Tyler Boone
2018-06-27 16:45:02
sustaining io on cloud offerings usually gets one throttled

Rotonen
2018-06-27 16:45:03
yup

Shoots
2018-06-27 16:45:22
using same config just commenting out the node_host and pool_host when I switch

Shoots
2018-06-27 16:48:16
over 25 min now
INFO: pre-caching snowfield: loaded 7 gb of 102 (6%)

Shoots
2018-06-27 16:48:51
normal memfield mining it takes about 10 minutes to load 185gb and use all of my memory

Shoots
2018-06-27 16:53:21
ok there we go, I was using PoolMiner that was built with bazel build :all, if I build and use PoolMiner_deploy.jar it loads faster

Shoots
2018-06-27 16:53:40
strange

Fireduck
2018-06-27 16:53:51
that was with a fresh git pull

Shoots
2018-06-27 16:54:59
i also have much better performance with the deployment builds

Rotonen
2018-06-27 16:55:15
that’s not expected?

Rotonen
2018-06-27 16:55:55
not at all

Fireduck
2018-06-27 16:56:09
should be no difference other than maybe a few hundred ms at startup

Fireduck
2018-06-27 16:56:24
unless someone fucked up the code in master

Fireduck
2018-06-27 16:56:33
implicit jvm flags via bazel? it’s a weird target class

Rotonen
2018-06-27 16:56:47
@Shoots if you want to pull down the 1.1.2 release tag and build that, I'd appreciate that

Fireduck
2018-06-27 16:56:51
it could be a performance regression

Fireduck
2018-06-27 16:57:29
@Rotonen good point. in that case, I'd expect better via bazel-bin/PoolMiner

Fireduck
2018-06-27 16:57:33
he was also noting earlier he had the best performance with 1.0.7

Rotonen
2018-06-27 16:58:29
yeah, might be a regression

Fireduck
2018-06-27 16:58:44
that entire thing needs to be restructured

Fireduck
2018-06-27 16:58:53
less emphasis on smoothly handling field changes

Fireduck
2018-06-27 16:58:56
i’ll do a benchmark sweep across tags later

Rotonen
2018-06-27 16:59:12
more emphasis on allowing field peices to have various optimization/buffer/behavior on or off

Fireduck
2018-06-27 16:59:19
@Rotonen I love you

Fireduck
2018-06-27 16:59:26
i’d just drop the whole field change stuff

Rotonen
2018-06-27 16:59:43
am I not pulling 1.1.2 when I do git pull?

Shoots
2018-06-27 16:59:45
the battery of hits and misses confuses people

Rotonen
2018-06-27 16:59:49
or at least kick it out to a base layer and restart everything

Fireduck
2018-06-27 18:36:45
1:20 min only 43gb loaded into memory

Shoots
2018-06-27 18:37:18
it started faster, but as soon as it started printing out that it was stalled it slowed right down

Shoots
2018-06-27 18:44:46
Just downloaded the 1.12 release and its loading fine, so it must be something to do with the version I was building

Shoots
2018-06-27 19:16:32
ok so now I just want to test to see if this version will go over my max amount of memory which is 185gb, but Im getting out of memory error when it hits 137gb.

Shoots
2018-06-27 19:16:33
java -Xmx185g -jar bazel-bin/PoolMiner_deploy.jar /var/snowblossom/configs/miner-mainnet.conf

Shoots
2018-06-27 19:34:07
define ’out of memory error’

Rotonen
2018-06-27 19:34:09
yeah super weird

Shoots
2018-06-27 19:34:10
Jun 27, 2018 7:33:01 PM snowblossom.miner.PoolMiner$MinerThread run
WARNING: Error: java.lang.OutOfMemoryError: Java heap space

Shoots
2018-06-27 19:34:33
only 137/185gb used, and Im using xmx flag to set the java limit higher

Shoots
2018-06-27 19:34:39

Shoots
2018-06-27 19:34:46
try with a matching -Xms

Rotonen
2018-06-27 19:35:10
do also note all java objects have overhead

Rotonen
2018-06-27 19:35:20
yes

Shoots
2018-06-27 19:35:29
but look at the htop screenshot

Shoots
2018-06-27 19:35:34
far from using the 185gb

Shoots
2018-06-27 19:35:56
never had this problem with other version, only since I built :all with 1.1.2 source

Shoots
2018-06-27 19:36:02
who's the crazy 4hk mining guy?

Clueless
2018-06-27 19:36:37
@Shoots i do not fully trust accounting from htop, use top ot atop

Rotonen
2018-06-27 19:37:07
htop has been accurate for me until now, Ill check top

Shoots
2018-06-27 19:37:47
i’ve had various silly issues with htop over the past 15 years

Rotonen
2018-06-27 20:00:10
well it eventually ran out of mem as well, not sure why 1.1.2 gives me so much lower hr than the older 1.0 version I was using before

Shoots
2018-06-27 20:00:18
3.4 > 2.5mh

Shoots
2018-06-27 20:01:00
probably terrible code by drunks

Fireduck
2018-06-27 20:01:25
lol

Shoots
2018-06-27 20:02:02
I have both the old version and the new version built on this same vm, I can run both and I get 3.4 on one and 2.5 on the other

Shoots
2018-06-27 20:02:26
I even built 1.1.2 from source cause someone mentioned doing that would help with threading

Shoots
2018-06-27 20:02:28
and this is memory mining?

Fireduck
2018-06-27 20:02:32
yes

Shoots
2018-06-27 20:02:48
would you mine filing an issue on github?

Fireduck
2018-06-27 20:02:48
Im going to try some other versions and test

Shoots
2018-06-27 20:03:27
sounds like something got broken when the pre-cache code went in, maybe

Fireduck
2018-06-27 20:09:46
Im trying 1.0.7 the first version with pre cache

Shoots
2018-06-27 20:28:31
yup 1.0.7 is slower too, 48c vm, gets 3.4mh on 1.0.6 and only 2.5mh on 1.0.7+

Shoots
2018-06-27 20:29:07
thats using the poolminer

Shoots
2018-06-27 20:29:15
sure

Fireduck
2018-06-27 23:08:35
@Shoots whats your hardware to get 3.4M/s?

stoner19
2018-06-27 23:08:45
compute instance?

stoner19
2018-06-27 23:08:49
48c vm

Shoots