ARM benchmarks for the erasure code plugin

ARM benchmarks for the erasure code plugin

[Loic Dachary]

[-- Attachment #1: Type: text/plain, Size: 525 bytes --]

Hi Sandon,

Would it be possible to get access to an ARM based machine to run erasure code plugins benchmarks ? The idea is to get a baseline similar to what https://github.com/ceph/ceph/pull/1875 shows and figure out where we stand. At this point any ARM processor will do. Eventually we will want to give NEON[1] enabled processors a try but right now we have nothing take advantage of that ;-)

Cheers

[1] http://en.wikipedia.org/wiki/NEON_%28instruction_set%29#NEON

-- 
Loïc Dachary, Artisan Logiciel Libre


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 263 bytes --]

Re: ARM benchmarks for the erasure code plugin

[Sage Weil]

The machine types are tala and saya; you should be able to grab one with 
teuthology-lock.  Hopefully they are up...

sage

On Fri, 27 Jun 2014, Loic Dachary wrote:

> Hi Sandon,
> 
> Would it be possible to get access to an ARM based machine to run erasure code plugins benchmarks ? The idea is to get a baseline similar to what https://github.com/ceph/ceph/pull/1875 shows and figure out where we stand. At this point any ARM processor will do. Eventually we will want to give NEON[1] enabled processors a try but right now we have nothing take advantage of that ;-)
> 
> Cheers
> 
> [1] http://en.wikipedia.org/wiki/NEON_%28instruction_set%29#NEON
> 
> -- 
> Lo?c Dachary, Artisan Logiciel Libre
> 
> 

Re: ARM benchmarks for the erasure code plugin

[Sandon Van Ness]

Yeah, they should be up and available as we are not testing them 
anymore. They do have issues where they randomly went down before and if 
any of them are down feel free to to hit me up and I can get them fixed up.

says37->saya44 are the newer machines that have the faster CPU's and 
more ram and stuff (newer arch).

On 06/27/2014 10:16 AM, Sage Weil wrote:
> The machine types are tala and saya; you should be able to grab one with
> teuthology-lock.  Hopefully they are up...
>
> sage
>
> On Fri, 27 Jun 2014, Loic Dachary wrote:
>
>> Hi Sandon,
>>
>> Would it be possible to get access to an ARM based machine to run erasure code plugins benchmarks ? The idea is to get a baseline similar to what https://github.com/ceph/ceph/pull/1875 shows and figure out where we stand. At this point any ARM processor will do. Eventually we will want to give NEON[1] enabled processors a try but right now we have nothing take advantage of that ;-)
>>
>> Cheers
>>
>> [1] http://en.wikipedia.org/wiki/NEON_%28instruction_set%29#NEON
>>
>> -- 
>> Lo?c Dachary, Artisan Logiciel Libre
>>
>>

Re: ARM benchmarks for the erasure code plugin

[Loic Dachary]

[-- Attachment #1: Type: text/plain, Size: 1686 bytes --]

Works ! Thanks :-)

$ ./virtualenv/bin/teuthology-lock --lock-many 1 --machine-type saya --owner loic@dachary.org
...

processor	: 3
model name	: ARMv7 Processor rev 0 (v7l)
BogoMIPS	: 2793.47
Features	: swp half thumb fastmult vfp edsp neon vfpv3 tls vfpd32
CPU implementer	: 0x41
CPU architecture: 7
CPU variant	: 0x3
CPU part	: 0xc09
CPU revision	: 0

Hardware	: Highbank
Revision	: 0000
Serial		: 0000000000000000
ubuntu@saya001:~$

On 27/06/2014 19:23, Sandon Van Ness wrote:
> Yeah, they should be up and available as we are not testing them anymore. They do have issues where they randomly went down before and if any of them are down feel free to to hit me up and I can get them fixed up.
> 
> says37->saya44 are the newer machines that have the faster CPU's and more ram and stuff (newer arch).
> 
> On 06/27/2014 10:16 AM, Sage Weil wrote:
>> The machine types are tala and saya; you should be able to grab one with
>> teuthology-lock.  Hopefully they are up...
>>
>> sage
>>
>> On Fri, 27 Jun 2014, Loic Dachary wrote:
>>
>>> Hi Sandon,
>>>
>>> Would it be possible to get access to an ARM based machine to run erasure code plugins benchmarks ? The idea is to get a baseline similar to what https://github.com/ceph/ceph/pull/1875 shows and figure out where we stand. At this point any ARM processor will do. Eventually we will want to give NEON[1] enabled processors a try but right now we have nothing take advantage of that ;-)
>>>
>>> Cheers
>>>
>>> [1] http://en.wikipedia.org/wiki/NEON_%28instruction_set%29#NEON
>>>
>>> -- 
>>> Lo?c Dachary, Artisan Logiciel Libre
>>>
>>>
> 

-- 
Loïc Dachary, Artisan Logiciel Libre


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 263 bytes --]

Re: ARM benchmarks for the erasure code plugin

[Loic Dachary]

[-- Attachment #1: Type: text/plain, Size: 1498 bytes --]



On 27/06/2014 19:23, Sandon Van Ness wrote:
> Yeah, they should be up and available as we are not testing them anymore. They do have issues where they randomly went down before and if any of them are down feel free to to hit me up and I can get them fixed up.
> 
> says37->saya44 are the newer machines that have the faster CPU's and more ram and stuff (newer arch).

Thanks for the tip. I'll beg Mark to lend me one ;-)

> 
> On 06/27/2014 10:16 AM, Sage Weil wrote:
>> The machine types are tala and saya; you should be able to grab one with
>> teuthology-lock.  Hopefully they are up...
>>
>> sage
>>
>> On Fri, 27 Jun 2014, Loic Dachary wrote:
>>
>>> Hi Sandon,
>>>
>>> Would it be possible to get access to an ARM based machine to run erasure code plugins benchmarks ? The idea is to get a baseline similar to what https://github.com/ceph/ceph/pull/1875 shows and figure out where we stand. At this point any ARM processor will do. Eventually we will want to give NEON[1] enabled processors a try but right now we have nothing take advantage of that ;-)
>>>
>>> Cheers
>>>
>>> [1] http://en.wikipedia.org/wiki/NEON_%28instruction_set%29#NEON
>>>
>>> -- 
>>> Lo?c Dachary, Artisan Logiciel Libre
>>>
>>>
> 
> -- 
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

-- 
Loïc Dachary, Artisan Logiciel Libre


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 263 bytes --]