You have already a raid-0 configuration, put SSD drives on it,
 you'd find some benchs here 
http://www.tomshardware.com/reviews/ssd-raid-iops,2848-6.html 

 Moses Odd

----- Original Message -----
From: Tom Hoar
Sent: 04/19/11 06:41 PM
To: Moses Odd
Subject: Re: [Moses-support] How much Ram for Europarl?

 Well, that's essentially my point. SSD's seem like a cost-effective 
alternative to expensive RAM and less complex (fewer parts) than RAID-0. I've 
had jobs, however, that generated over 150 GB in temp files and ran out of 
space (not fun). So, an 80 GB SSD disk isn't big enough. That's why I was 
curious about the configuration. 
 Also, we run binarized tables and language models (KenLM) from a RAID-0, which 
recovers 50-75% of the lost performance compared to memory-loaded text tables 
and models. It would seem like a fast SSD over iodrive/PCIe would shift the 
bottleneck to the IO-to-memory bus. I don't see much advantage to installing 
all of Linux on SSD over well-planned and configured mounts. Then again, a 
server configuration is  under 4 GB, so maybe a complete install is just 
easier. 
 Tom
 On Tue, 19 Apr 2011 16:16:49 +0000, "Moses Odd" < moses4o...@gmx.com > wrote:

Hi,

 ideally swap and tmp have to stay in fast disks,
 why not in ram disk or SSD ?
 See this http://en.wikipedia.org/wiki/File:Ddrdrive_x1.jpg 
http://en.wikipedia.org/wiki/Solid-state_drive 

 some commercial producer 
http://www.fusionio.com/products/iodrive/ 
http://www.fusionio.com/products/iodriveduo/ 
http://www.ocztechnology.com/ocz-revodrive-pci-express-ssd.html 
http://www.ramsan.com/products 
http://www.3dnow.net/phpBB2/viewtopic.php?f=1&t=1649 

 with a SSD you can reach 200~400 MB/s rw bandwidth with ~1 ms access time on 
Sata
 with a iodrive you can reach 750~1500 MB/s rw bandwidth with ~30 µs access time

 where an HD you can reach 50~70 MB/s rw bandwidth with 12~25 ms access time on 
Sata

 For the 80 GB HD, I defined a 16 GB swap partition and the rest for /tmp.
 Could be interesting to install also Linux on SSD.

 Moses 

----- Original Message -----
From: Tom Hoar
Sent: 04/19/11 03:08 PM
To:  maria.mach...@ec.europa.eu 
Subject: Re: [Moses-support] How much Ram for Europarl?

 Wow. Excellent tip. After looking at pricing on Ebay, it's well-worth the 
$200-$500 investment. We run multiple swap partitions across 4 physical disks, 
plus a RAID 0 for /tmp space. It speeds things up, but not to your extent. Can 
you share some details about the drive specs, i.e. interface, cache size, etc? 
Also, what is your configuration between swap and tmp size? Tom On Mon, 18 Apr 
2011 18:01:17 +0200, < maria.mach...@ec.europa.eu > wrote: > Hello, > > 
Building the phrase table really used to take me a long, long time. > > I have 
a 4-processor computer with 8 GB RAM and with a 12 million > segment corpus 
(about 0.5 billion words EN+PT), the whole training > took about 7 days, of 
which 2 days to build the phrase table (using > the swap too). > > However, now 
I have a 80 GB solid-state drive installed for the swap > and temp files and 
the training of a larger corpus (14 million > segments) took about the same 
time. The main difference was in the > building of the phrase !
 table: it took only 7 hours. Beautiful! > > I hope this information may be 
useful to you ... although the corpus > you want to train is not as large. > > 
Maria José > > -----Original Message----- > From:  
moses-support-boun...@mit.edu  > [mailto: moses-support-boun...@mit.edu ] On 
Behalf Of Tom Hoar > Sent: Monday, April 18, 2011 4:05 PM > To: David Wilkinson 
> Cc:  moses-support@mit.edu  > Subject: Re: [Moses-support] How much Ram for 
Europarl? > > > Your report of 100% physical usage, growing swap usage and low 
CPU > load > is normal when working with limited RAM machines. With only 4 Gb 
Ram > and > the new (larger) EuroParl v6 corpus, you could train for 3 or 4 
days > depending on how you setup your swap partition. Even then, it's > 
possible > you will run out of RAM before it's finished. Upgrading to 8 Gb ram 
> is a > move in the right direction. > > Once it's finished training, you'll 
want to use the binarized the > tables and language model, which MMM's 
train-1.11 sc!
 ript creates. > > Tom > > > On Mon, 18 Apr 2011 14:52:10 +0100, Philip
p Koehn > < pko...@inf.ed.ac.uk > > wrote: >> Hi, >> >> I am not familiar with 
the MMM setup, but one of the causes >> of memory use may be the translation 
table. You should use >> the on-disk translation table. >> >> -phi >> >> On 
Mon, Apr 18, 2011 at 2:47 PM, David Wilkinson >> < davidzw...@hotmail.com > 
wrote: >>> I have set up an Ubuntu 10.04 system with the >>> 
moses-for-mere-mortals >>> scripts. The default corpus trained in about 6-7 
hours on my system >>> (Athlon >>> x3 3.2Ghz, 4Gb Ram). I am now trying to 
train the system with the >>> Europarl >>> German-English parallel corpus 
(about 45m words in each language), >>> again >>> using the default 
moses-for-mere-mortals settings. The system has >>> been >>> running for 24 hrs 
and is currently using all the physical memory >>> and about >>> 1.2Gb of swap. 
None of the cores are being used more than 10%, so >>> like this >>> it will 
take a very long time to finish. If I double the ram to >>> 8gb, >>> will >>> 
this be suff!
 icient? >>> Many Thanks >>> David >>> 
_______________________________________________ >>> Moses-support mailing list 
>>>  Moses-support@mit.edu  >>> 
http://mailman.mit.edu/mailman/listinfo/moses-support  >>> >>> >> 
_______________________________________________ >> Moses-support mailing list 
>>  Moses-support@mit.edu  >> 
http://mailman.mit.edu/mailman/listinfo/moses-support  > > 
_______________________________________________ > Moses-support mailing list >  
Moses-support@mit.edu  > http://mailman.mit.edu/mailman/listinfo/moses-support  
_______________________________________________ Moses-support mailing list  
Moses-support@mit.edu http://mailman.mit.edu/mailman/listinfo/moses-support
_______________________________________________
Moses-support mailing list
Moses-support@mit.edu
http://mailman.mit.edu/mailman/listinfo/moses-support

Reply via email to