staff blogs

distributed.net staff keep (relatively) up-to-date logs of their activities in .plan files. These were traditionally available via finger, but we've put them on the web for easier consumption.

2005-09-14

nugget [14-Sep-2005 @ 16:08]

Filed under: Uncategorized — nugget @ 16:08 +00:00

:: 14-Sep-2005 16:08 GMT (Wednesday) ::

A moment of silence for oldnodezero…

859 days ago (check my .plan) the UPS sitting behind our server
oldnodezero.distributed.net exploded in an exciting boom described even
today as “the great UPS cataclysm” by visi.com staff. The UPS explosion
dashed the spiffy 761 day uptime on the server.

Today the box finally died, setting a new record 859 day uptime. Attempts
to resuscitate the box have not been successful. It died with page fault
warnings on the console and now it won’t even POST. ODD is going to swing
by visi.com tomorrow to pick up the carcass and ship it down here to Austin
for forensics.

It really says a lot about a colo facility when they can provide that kind
of stability though. Visi has been babysitting this box since we put it
into production in 1998. It’s the first piece of hardware that we ever
bought for distributed.net and it’s been in productive use ever since.

It’s an AT case. Pre-ATX. An AMD K6 with a gigantic 8GB IDE hard drive.
Garage-built consumer junk, but it sure did hold up well. We’ll be
replacing it with a more modern 1U rackmount machine soon. In the
meantime though, we’re down to just the single server for dns and web.

Thanks again to visi.com who have consistently provided peerless colo
services. I’m sure if we manage to beat 859 days this time around, the
eventual failure will again be something beyond their control.

2005-09-13

bovine [13-Sep-2005 @ 16:54]

Filed under: keyservers — bovine @ 16:54 +00:00

:: 13-Sep-2005 16:54 GMT (Tuesday) ::

Last week we added a new Aussie keyserver, located in Brisbane.
Clients have already begun to automatically connect to it and it
is performing great so far.

Our secondary webserver (one of two servers that host www.d.net) is
currently offline. Since all of the content it was hosting was
mirrored, this should not have much immediate impact on us. We’ve
adjusted our load-balancing to exclude it, although there may be some
delays until the DNS changes fully propagate. (That machine had been
online for nearly 8 years and would have had an uptime of 858 days at
the time it went offline.)

2005-08-01

decibel [01-Aug-2005 @ 11:40]

Filed under: stats — decibel @ 11:40 +00:00

:: 01-Aug-2005 11:40 GMT (Monday) ::

Stats were offline for approximately 8 hours due to a network outage. They’re
back online now. Sorry for the inconvenience.

2005-07-06

snikkel [06-Jul-2005 @ 16:06]

Filed under: clients — snikkel @ 16:06 +00:00

:: 06-Jul-2005 16:06 GMT (Wednesday) ::

A new version of the client source has been posted. It includes
all the core improvements introduced with the recent clients.
We’ve added the MA4 core which targets Pentium 4 chips, but is surpassed
in speed by other cores, the MMX core which targets original Pentium era
chips and the KBE-64 core which is optimized for 64-bit AMD K8 and Pentium
4 chips. In addition we applied speed improvements from contributors for
the DG3 and GO2 cores. We also corrected issues in DG3, SGP3 and SNJL
cores. In the OGR world, we’ve added a new assembly version of the x86
core which improves performance by around 30%.
Get the source at http://www.distributed.net/source/.

2005-06-30

bovine [30-Jun-2005 @ 06:41]

Filed under: clients,project status — bovine @ 06:41 +00:00

:: 30-Jun-2005 06:41 GMT (Thursday) ::

New clients have been released and can be downloaded from
http://www.distributed.net/download/ Users running any previous x86
client versions are encouraged to update.

Included with client version 2.9010.495 and higher are some exciting
speed improvements. Three new RC5-72 cores have been added, including
a new 64-bit core for AMD64 and two existing x86 RC5-72 cores have
been sped up. In addition a new OGR core has been added that provides
an approximate 30% improvement. We are eternally grateful for our
contributors’ ingenuity.

OGR25-P2 is progressing well. We have now checked over 47.66% of all
work in the first of nine sub-phases. Overall we have completed 7.66%
of OGR25-P2. For more information about the sub-phases look here:
http://faq.distributed.net/cache/230.html Note that due to an
oversight the last time we updated OGR completion on stats, it has
been showing the percent completion of the first sub-phase. We will be
going back to displaying overall completion, since this is what most
users are interested in.

A few minor issues with the clients have been corrected as well. The
‘dg3’, ‘sgp3’, and ‘snjl’ cores present in versions prior to
2.9010.495 miscalculate a small number of predictable keys in every
block, rendering some blocks containing those keys discarded for
stats. These cores were used as the default choice for 32-bit clients
running on Cyrix Cx486, AMD K5, Via C3 Nehemiah, Pentium 4, Pentium M
and 64-bit clients running on the AMD64/Intel EM64T processors.

Note that at most only 2% of the calculated keys within each block
processed by those cores are affected. This won’t affect the normal
running of the project in any way as we’ll re-process these keys if
needed.

All current x86 clients need to be upgraded to prevent the amount of
work that would need to be recalculated from growing. This also means
that all RC5-72 work from x86 clients v2.9001.474 up to v2.9009.494
(all clients released to date) will stopped being credited at a future
date.

In the past few months, a growing minority of participants have chosen
to try to cheat their way to higher stats rankings by making binary
alterations to our clients. Fortunately, great progress has been made
on both removing the largest users of these hacked clients and
implementing automatic filtering techniques to block future tainted
work.

Cheaters are removed from stats manually after intensive study of
their flushing patterns, to prevent users that tested such clients and
accidentally submitted a few blocks from being removed. Due to the
amount of manual work involved with this, a few smaller accounts may
still face removal in the near future.

We thank you for all of your contributions and hope that you will
stick with us into the future as we continue to make history. :)

2005-06-15

bovine [15-Jun-2005 @ 17:15]

Filed under: clients — bovine @ 17:15 +00:00

:: 15-Jun-2005 17:15 GMT (Wednesday) ::

The previous v2.9010-495 release candidates (on x86 platforms only)
have been withdrawn and not officially released, due to some minor
performance and some inaccurate auto core-selection. (Thanks to those
who participated in the testing and identification of those problems.)

A few new v2.9011-496 release candidates have already been added to
the pre-release page. Eager users interested in helping provide test
coverage and report any issues for these new sets of clients can
download them from: http://www.distributed.net/download/prerelease.php

If no additional blocking problems are discovered, we hope to
officially release these clients for general consumption within the
next week or so.

2005-05-18

bovine [18-May-2005 @ 16:35]

Filed under: clients — bovine @ 16:35 +00:00

:: 18-May-2005 16:35 GMT (Wednesday) ::

I have just uploaded revised v2.9010.495b release candidates for Win32
to fix an unexpected core selection problem on AMD Athlon 64 FX or
newer processors. If you had already downloaded the previous Win32
prerelease client already, it is recommended that you download the
updated one.

Pre-release Linux clients for x86 and AMD64 have also now available.

As before, we’d like to recommend that only users who are willing to
perform early testing and open bug reports download these clients now.

http://www1.distributed.net/download/prerelease.php

2005-05-17

bovine [17-May-2005 @ 02:40]

Filed under: keyservers,stats — bovine @ 02:40 +00:00

:: 17-May-2005 02:40 GMT (Tuesday) ::

There was some unexpected server hardware downtime today, which has
led to some backlog accumulating and a delay in the start of today’s
daily stats processing. We’re currently processing the backlog right
now and hope to catch up within the next few hours. It is not
expected that any significant amount of work was lost, due to the
buffering nature of our network.

However please be aware that the statsrun for May 16 (once it has
finished processing) may be a little low, since most of the backlog
will be counted in tomorrow’s stats.

bovine [17-May-2005 @ 02:28]

Filed under: clients — bovine @ 02:28 +00:00

:: 17-May-2005 02:28 GMT (Tuesday) ::

I’ve just put a few release candidate clients for v2.9010.495 on the
pre-release page. At this point, we’d like to recommend that only
users who are willing to perform early testing and open bug reports
download these clients now. We hope to formally release these clients
later this week.

http://www1.distributed.net/download/prerelease.php

2005-05-06

decibel [06-May-2005 @ 17:12]

Filed under: stats — decibel @ 17:12 +00:00

:: 06-May-2005 17:12 GMT (Friday) ::

I’ve finally fixed the stats bug, and stats are caught up again (just in time
for statsrun…)

« Newer Posts — Older Posts »