MS-DOS client

[flippinfleck]

I am trying to figure out how to set up an old 486dx4 100 and the moo cow. I have downloaded the proper client for this, but am a little confused as to how to "fetch and flush via email". It is asking me for a preferred blocksize. What is this blocksize it is asking me for? It wants "numblocks=yyyy" also. Can anyone help me?

 

[Polo]

2^28 keys = 1 Work Unit.
1 block=1 packet= 2^XX keys where 27<XX<34.
1 2^28 packet= 1 WU
2^29 = 2 WUs
2^30 = 4 WUs
2^31 = 8 WUs
2^32 = 16 WUs
2^33 = 32 WUs
Your client can fetch/flush different size of packets/blocks. You can choose a preferred blocksize for your current client to match the cracking speed. For your 486DX4 100, according to Mika's speed page, your max speed is about 98Kkeys/s. If your machine is on and cracking 24/7 then choose a 2^32 size of packet and fetch/fluch once a day. If you set it at 2^33, you won't see results on a daily basis.
Then it will crack 31 WUs per day and fetch/flush packets containing 16 WUs.
Set the number of blocks numblocks=10 and your in buffer will contain 10 packets of 16 WUs (160 WUs). That's enough to last 5 days without fetching. Set it to what you think is good for you... 😉

Hope this helps.

 

[Sukhoi]

I don't know this for sure, but I think I remember hearing that the MS-DOS client is kinda slow. Does anyone know for sure?

Can't you run Windows 95 on the machine? I have an AMD 486 DX4-100 running Win95A, and I get about 94 Kkeys/sec. 🙂 If not, can you run Klinux on the machine?

 

[phule]

The MS-DOS client is no &quot;slower&quot; than any other client afaik. It should actually be as faster because there's nothing it has to share the processor with. Of course the fact that you're using a 486/100 means that none of this really matters...but anyway.

 

[phule]

sending a blank email to fetch@distributed.net returns:
************************************************************
A complete request was not found. At the very minimum, your request should specify the number of blocks that you would like, via the 'numblocks'keyword.

INSTRUCTIONS FOLLOW:
This message has been sent to you because you sent mail
to fetch@distributed.net. The attached is the output of
&quot;dnetc -fetch&quot;.

Include &quot;numblocks=yyyy&quot; anywhere in the body of
your message. Note that the client may impose an
upperlimit of the number of workunits you can
request in 1 fetch.

To specify a preferred blocksize, include
&quot;blocksize=xx&quot; anywhere in the body of your mesasage,
&quot;xx&quot; being between 28 and 33.

To request OGR keys, include &quot;contest=OGR&quot; anywhere in
the body of your message. Default is RC5.

Other than these flags, the contents of any messages sent
to fetch@distributed.net are ignored.

The attached buffers contain approximately the number of
workunits you requested by the keyword &quot;numblocks&quot;
Note that now, numblocks does not indicate number of blocks
of packets, but workunits. This can mean you actually get
less packets/blocks, since blocks can contain multiple
workunits. This makes the behavior of the keyword
&quot;blocksize&quot; a little different, since this this keyword
doesn't influence the number of workunits you get.

Three &quot;-fetch&quot; attempts are made, in an attempt to
overcome any network errors.

 

[Wellcky]

nevermind😱

 

[flippinfleck]

Thanks for all the help! Yeah, I know it is slow as all getout, but every little bit counts! And besides, what better reason to have a motherboard sitting outside of a case in a corner? I would throw win95 on it, but somehow lost my copy 🙁