Thanks Steve, for some reason the spreadsheet actually has
HEPSPEC*1000/4 rather then just HEPSEPC*250. No idea why that is.
Any other changes before I send the revised report to Jeremy this afternoon?
Cheers,
Matt
On 10/17/2014 05:00 PM, Stephen Jones wrote:
> I do believe that SI2K = HEPSPEC * 250
>
> (I call them "bogoSI2K" BTW, to distinguish them from the real SI2K
> benchmark/ )So Sheffield cpu power would equate to 12.2 * 250 * 1064 =
> 3245200.0
>
> Steve
>
>
>
>
>> Hi Matt
>>
>> We have
>>
>> cores: CE_LOGCPU=1064
>> Benchmark 12.20-HEP-SPEC06
>>
>> Thanks
>> Elena
>>
>> On 17 Oct 2014, at 12:18, Matt Doidge <[log in to unmask]> wrote:
>>
>>> Thanks for the update everyone - I knew gstat was on it's away out but I
>>> didn't realise how rubbish it had become.
>>>
>>> Elena, rebas and ggus are telling lies could you please send me the
>>> right figures, like Steve did for Liverpool.
>>>
>>> Thanks,
>>> Matt
>>>
>>> On 10/17/2014 11:49 AM, Elena Korolkova wrote:
>>>> Hello,
>>>>
>>>> by the way rebus gives incorrect numbers as well:
>>>>
>>>> http://wlcg-rebus.cern.ch/apps/capacities/sites/.
>>>>
>>>> I changed the numbers recently (as we added resources) but neither
>>>> rebus or gstat reflect the change.
>>>>
>>>> Elena
>>>>
>>>>
>>>> On 17 Oct 2014, at 11:20, Stephen Jones <[log in to unmask]>
>>>> wrote:
>>>>
>>>>> Hi Matt,
>>>>>
>>>>> I believed that gstat is useless. There was something on TB_SUPPORT
>>>>> about
>>>>> it, I thought.
>>>>>
>>>>> Anyway, the "proper" values for Liverpool are:
>>>>>
>>>>>> SI2K
>>>>>> Q2 Q3
>>>>>> LIV 4337456 4336880
>>>>>>
>>>>>> CORES
>>>>>>
>>>>>> LIV 1384 1384
>>>>>
>>>>> I assume the 0.013% difference is a calculation error as the cores
>>>>> have
>>>>> not changed.
>>>>>
>>>>> Steve
>>>>>
>>>>>
>>>>>
>>>>>> Hello Everybody,
>>>>>> Thanks to Steve for getting the Liverpool Narrative to me on time.
>>>>>> I'm
>>>>>> still putting together the report - as you know one of the columns is
>>>>>> the gstat CPUs and SI2K for sites, to be obtained from gstat.
>>>>>>
>>>>>> But the numbers in gstat at the moment are lower then those from Q3
>>>>>> for
>>>>>> everyone except Lancaster- is this expected? Perhaps an ARC CE
>>>>>> publishing oddity? Or have sites been retiring kit? Or is gstat just
>>>>>> being wrong?
>>>>>>
>>>>>> SI2K
>>>>>> Q2 Q3
>>>>>> MAN 9666169 8920480
>>>>>> LIV 4337456 2400960
>>>>>> SHEF 2818080 1409040
>>>>>>
>>>>>>
>>>>>> CORES
>>>>>>
>>>>>> MAN 3958 3646
>>>>>> LIV 1384 732
>>>>>> SHEF 912 456
>>>>>>
>>>>>> (As you can see the CORES/SI2K proportions seem to remain the same).
>>>>>>
>>>>>> Cheers,
>>>>>> Matt
>>>>>>
>>>>>> On 10/01/2014 10:31 AM, Matt Doidge wrote:
>>>>>>> Hello all,
>>>>>>> Well Q3 disappeared in a flash - and it's quarterly report time
>>>>>>> again!
>>>>>>> That means narratives need to be written. Could sites please get
>>>>>>> their
>>>>>>> narratives to me by the 14th of October? That will give us a few
>>>>>>> days to
>>>>>>> iron out any wrinkles before the report has to be in.
>>>>>>>
>>>>>>> Thanks!
>>>>>>> Matt
>>>>>>
>>
|