[Halld-offline] Hall-D Computing plan and storage of simulated data.

[Graham Heyes]

Chip,
    Yes, that is planned but the machines are not fully operational yet and I want a first cut at updated numbers before the March review. 
     Graham

On Feb 11, 2011, at 9:11 PM, Chip Watson <watson at jlab.org> wrote:

> Please also plan to normalize to both the recent 8 core Intel boxes (dual quad), and the 32 core AMD boxes (quad 4+4) that are now being installed.
> 
> Graham Heyes wrote:
>> As Sandy pointed out. I have a typo in my last mail. I meant 64-bit nodes of course. 
>> Graha
>> 
>> On Feb 11, 2011, at 3:37 PM, Graham Heyes <heyes at jlab.org> wrote:
>> 
>>  
>>> Curtis,
>>>   Chip asked for SpecInt numbers. This time around I am asking the questions and am asking for "time to process one event" normalized to out current generation 32-bit farm nodes. This should be much easier to measure/estimate. I have spreadsheets that can take those values and estimate performance in terms of standardized benchmarks and then estimate how many nodes we need.
>>> 
>>>   Regards to all,
>>>       Graham
>>> 
>>> On Feb 11, 2011, at 3:27 PM, Curtis A. Meyer wrote:
>>> 
>>>