[Bioclusters] Need some advice on a cluster for EST/cDNA assembly, clustering

[Joe Landman]

Hi Elia:

   1 GB sticks (good ones) are about $180-200 US  ( 1GB DDR400 ECC ) for 
Opterons.  2 GB sticks are a bit more, about $500 - 550 or so.

   32 GB would be 16 of these sticks, or about $8000US.  16 GB would be 
8 of these sticks or about $4000US.  There is something oddly familiar 
about paying more for the memory than the rest of the box, reminiscent 
of Cray days ...

   You could always opt for slower memory, and get DDR266 to drop the 
price per stick by 20% or so.

   From what we have seen, the lower end of machines going in to 
customers typically have 2 GB ram on them, in 512MB stick form 
(currently the best  price capacity, because there are so many of them). 
   We have built a few 16 GB machines for customers, and just quoted out 
a few 32 GB units.  You will be paying for memory, and there is a bit of 
a premium for the higher capacities.

   Future revs of publically available Opteron boards will be able to go 
a bit higher than 32 GB (8GB/CPU).  IBM's units are based on an MSI 
motherboard, and last I checked, IBM still sees a limited role for 
Opteron's in it's lineup.  Sun is offering 2 way and 4 way today.  Same 
with HP.  Some of the smaller vendors are offering real 8-way units 
today (2 Iwill or Arima MB's joined with a hypertransport HTX link or 
similar).  Very interesting stuff is coming in the near future from a 
number of vendors in this space.  Saw some amazing stuff at SC04 (not 
the stuff on the show floor).

Joe

Elia Stupka wrote:
> I've been hoping to be able to access large amounts of memory on 
> affordable servers for a while, in reality, though, the 4GB OS limit has 
> hardly been the issue since unfortunately the cost of memory is still 
> very high and hardware vendors seldom offer more than 4GBs per processor.
> 
> The Sun Opterons are the only ones (among the mainstream vendors) that 
> offered us a 4-way option with 32 GBs. The Apple G5s are still limited 
> to 8GB (4 per processor, probably when Tiger will be truly released they 
> will finally offer more memory slots?), IBM Opterons offer 16GB (still 
> only 4GB per processor), the blade versions are always limited in 
> memory, etc... then you are left with the usual suspects (Power5s, etc.) 
> who have been dealing with more memory for a long time, but at a nasty 
> price...
> 
> ...as long as it costs more to equip hardware with good amounts of 
> memory than it costs to buy the hardware, the refinement of 64-bit OS 
> for access to large amounts of memory can't take off properly, can it?
> 
> Cheers,
> 
> Elia
> 
> 
> On 27 Feb 2005, at 15:40, Joe Landman wrote:
> 
>>
>>
>> LAI Loong Fong wrote:
>>
>>> The process in OSX 10.3 can only see up to 4GB of ram. From what I
>>> understand, 10.4 may or may not resolve this issue. I did heard about 
>>> people
>>> buying from Apple but run YDL instead.
>>
>>
>> The big issue we keep seeing in 64 linux is running into ABI 
>> differences.  There are a few distributions that try to create a 
>> "pure" 64 bit or a mixed 64 bit environment, and neither do a perfect 
>> job. Usually you see this in terms of broken software 
>> installation/builds. Other times libraries are broken, or headers, or 
>> paths to the libraries/headers ...
>>
>> It would help if there were simple methods of tying paths to ABI's.
>>
>> In terms of the 4GB limit, this is an annoyance today, and I suspect 
>> for most folks, will become a major issue in short order.  I think YDL 
>> is a RH variant, so hopefully it has support built in (if it was built 
>> from the 64 bit ports of RH).  If not, it would be nice if OSX came in 
>> an OSX64 (if not today, then soon).
>>
>> I wonder if that would get compiler support though (the IBM compiler 
>> specifically).
>>
>> Joe
>>
>>> LAI Loong-Fong
>>> _______________________________________________
>>> Bioclusters maillist  -  Bioclusters at bioinformatics.org
>>> https://bioinformatics.org/mailman/listinfo/bioclusters
>>
>>
>> -- 
>> Joseph Landman, Ph.D
>> Scalable Informatics LLC,
>> email: landman at scalableinformatics.com
>> web  : http://www.scalableinformatics.com
>> phone: +1 734 786 8423
>>
>> _______________________________________________
>> Bioclusters maillist  -  Bioclusters at bioinformatics.org
>> https://bioinformatics.org/mailman/listinfo/bioclusters
> 
> 
> _______________________________________________
> Bioclusters maillist  -  Bioclusters at bioinformatics.org
> https://bioinformatics.org/mailman/listinfo/bioclusters

-- 
Joseph Landman, Ph.D
Founder and CEO
Scalable Informatics LLC,
email: landman at scalableinformatics.com
web  : http://www.scalableinformatics.com
phone: +1 734 786 8423
fax  : +1 734 786 8452
cell : +1 734 612 4615