these could be interesting as well: https://www.anandtech.com/print/13047/ngd-launches-catalina-2-programmable-ssds NGD Launches Catalina 2 Programmable SSDs: 16 TB - 32 TB, ARM A-53 Cores
It doesn't look like they are Keen on selling things
Cool concept
Contact sales if interested = it costs too much money so we need to hard sell it
What do you think of that rig for a memory machine as part of the distributed memory mining?
for anyone who downloaded the chunked field 8 torrent, it included an extra file and has been fixed
I suggest removing that torrent and getting the new one
it should recheck and resume with what you have
@Fireduck is that really a X399 in there? X390 would seem more plausible as the X399 is a naples dual socket platform
@Fireduck but if i parse block diagrams right, 8 hyperthreads per memory channel, seems legit
@Fireduck also maybe get a suitable case and a closed loop liquid cooler, and get a platinum or titanium certified seasonic power supply - never skimp on the PSU
`94% efficiency at 50% system load` seems nice, also fanless
@Fireduck i've not seen any performance hits or gains across miner versions, i suppose jvm tweaking and systems tweaking sidestep whatever it is others hit
I think the diff in hr from versions depends on the amount of cores/mem. I definitely see a difference on AWS and so do others that I've talked to.
also how exactly the memory they provision to you maps to the backing hardware
if it spreads across more channels, you win
snowplough is now running 1.2.0
7200 snow are generated per day?
144*50.
So yeah
@Rotonen I think the X399 is the motherboard chipset, but I don't know. I buy new parts infrequently enough where I am never really sure.
thanks, I'll put in that power supply in case I ever get that system
I ordered a refurbished Dell R900 server with 256gb DDR2
I'm sure it will sound like a jet engine
but at $700, hard to pass up
@Rotonen thanks for doing that testing across miner versions
in example for arktika it says "java -Xmx10g -jar Arktika_deploy.jar $1".
but there is no Arktika_deploy.jar.
you have to make it
bazel build :Arktika_deploy.jar
ah. oki. thanks.
What setup are you trying to do?
just curious
ssdhdd.conf. just for testing.
but i was getting an error: INFO: Error talking to mining pool: io.grpc.StatusRuntimeException: UNAVAILABLE: Channel shutdownNow invoked
so i decided to drink beer and watch football instead.
go belgium :fries:
so that error can occur if it isn't hashing and thinks it should maybe reset the work channel
might not be a problem
brazil out. lol.
^^ any idea?
how many threads on those fields?
layers
found zero chunks is also ungood
do you have the chunked data files?
ah. ok. i am using the normal snowfield.
you can either use split.sh or download the chunked one
i need the chunk files?
if you already have it downloaded, split.sh will be a lot faster
oki. i will try.
it is in tools
you'll have to run it in your snowblossom.7 directory
okay