All of lore.kernel.org
 help / color / mirror / Atom feed
* Re: [SPDK] Performance
@ 2016-12-06 16:58 Walker, Benjamin
  0 siblings, 0 replies; 12+ messages in thread
From: Walker, Benjamin @ 2016-12-06 16:58 UTC (permalink / raw)
  To: spdk

[-- Attachment #1: Type: text/plain, Size: 2585 bytes --]

On Tue, 2016-12-06 at 16:35 +0000, Harris, James R wrote:
> 
> > On Dec 6, 2016, at 5:49 AM, Kumaraparameshwaran Rathnavel <krath(a)cloudsimple
> > .com> wrote:
> > 
> > The setup is the default setup of NVMf target with Direct attached
> > Subsystem. The FIO plugin was at the target side and the IOPS from SPDK was
> > found to be around 2.5K and the IOPS from kernel was found to be 100k.
> 
> 
> Hi Param,
> 
> Can you check the AcceptorPollRate setting in your configuration file?  There
> was a recent fix to the default value for this parameter.  You should set it
> to at least 10000 (10ms).
> 
> https://github.com/spdk/spdk/commit/5f3761cf7ec50cf03183203dc50657a9f2a18908
> 
> Thanks,

Just to reinforce here - this is the solution. We saw this internally a few
weeks ago. Increasing the AcceptorPollRate to at least 10ms (you can go as much
as 1 second if you want, but the performance benefit decreases the larger you
make it), or moving the accceptor to a different core using the AcceptorCore
option will fix the problem.

> 
> -Jim
> 
> > 
> > Regards,
> > Param.
> > > On 06-Dec-2016, at 5:57 PM, Andrey Kuzmin <andrey.v.kuzmin(a)gmail.com>
> > > wrote:
> > > 
> > > 
> > > 
> > > On Tue, Dec 6, 2016, 15:24 Kumaraparameshwaran Rathnavel <krath(a)cloudsimpl
> > > e.com> wrote:
> > > > Hi All,
> > > > 
> > > > Is there any benchmarking done to prove that SPDK is better than the
> > > > Kernel driver?
> > > > 
> > > > We did run the sample driver provided from the SPDK and compared with
> > > > the kernel interface and found that SPDK is very slow. We did run FIO.
> > > > 
> > > 
> > > It's worth publishing your setup and results first.
> > > 
> > > Regards,
> > > Andrey
> > > > 
> > > > Can you describe the setup for the performance measurement between
> > > > Kernel and SPDK driver.
> > > > 
> > > > Regards,
> > > > Param.
> > > > _______________________________________________
> > > > SPDK mailing list
> > > > SPDK(a)lists.01.org
> > > > https://lists.01.org/mailman/listinfo/spdk
> > > > 
> > > -- 
> > > Regards,
> > > Andrey
> > > _______________________________________________
> > > SPDK mailing list
> > > SPDK(a)lists.01.org
> > > https://lists.01.org/mailman/listinfo/spdk
> > 
> > _______________________________________________
> > SPDK mailing list
> > SPDK(a)lists.01.org
> > https://lists.01.org/mailman/listinfo/spdk
> 
> _______________________________________________
> SPDK mailing list
> SPDK(a)lists.01.org
> https://lists.01.org/mailman/listinfo/spdk

^ permalink raw reply	[flat|nested] 12+ messages in thread
* Re: [SPDK] Performance
@ 2017-01-18 23:41 Walker, Benjamin
  0 siblings, 0 replies; 12+ messages in thread
From: Walker, Benjamin @ 2017-01-18 23:41 UTC (permalink / raw)
  To: spdk

[-- Attachment #1: Type: text/plain, Size: 1800 bytes --]

On Tue, 2016-12-06 at 16:58 +0000, Walker, Benjamin wrote:
> On Tue, 2016-12-06 at 16:35 +0000, Harris, James R wrote:
> > 
> > > On Dec 6, 2016, at 5:49 AM, Kumaraparameshwaran Rathnavel <krath@
> > > cloudsimple
> > > .com> wrote:
> > > 
> > > The setup is the default setup of NVMf target with Direct
> > > attached
> > > Subsystem. The FIO plugin was at the target side and the IOPS
> > > from SPDK was
> > > found to be around 2.5K and the IOPS from kernel was found to be
> > > 100k.
> > 
> > 
> > Hi Param,
> > 
> > Can you check the AcceptorPollRate setting in your configuration
> > file?  There
> > was a recent fix to the default value for this parameter.  You
> > should set it
> > to at least 10000 (10ms).
> > 
> > https://github.com/spdk/spdk/commit/5f3761cf7ec50cf03183203dc50657a
> > 9f2a18908
> > 
> > Thanks,
> 
> Just to reinforce here - this is the solution. We saw this internally
> a few
> weeks ago. Increasing the AcceptorPollRate to at least 10ms (you can
> go as much
> as 1 second if you want, but the performance benefit decreases the
> larger you
> make it), or moving the accceptor to a different core using the
> AcceptorCore
> option will fix the problem.

To follow up on this - we've now root caused exactly why our acceptor
polling function was slowing down the system so much. Before, I had
suggested polling less often or doing the polling on a separate core.
After these two commits

https://github.com/spdk/spdk/commit/d2c0feac8a13944bc9af7c60ed047809cf1f5d9dhttps://github.com/spdk/spdk/commit/4a95a81e694c4f58031b181a6127583d6f55483e

that is no longer necessary. Any remotely reasonable value for AcceptorPollRate
and placing the Acceptor on any core will now yield excellent performance
results with SPDK.

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 3274 bytes --]

^ permalink raw reply	[flat|nested] 12+ messages in thread
* Re: [SPDK] Performance
@ 2016-12-07  1:22 Yang, Ziye
  0 siblings, 0 replies; 12+ messages in thread
From: Yang, Ziye @ 2016-12-07  1:22 UTC (permalink / raw)
  To: spdk

[-- Attachment #1: Type: text/plain, Size: 1697 bytes --]

Hi Param,

For NVMf target, you can change the polling rate of nvmf target from 1 ms to 10 ms, it will improve the performance. It means that you can configure the AcceptorPollRate in the configuration file.

From: SPDK [mailto:spdk-bounces(a)lists.01.org] On Behalf Of Kumaraparameshwaran Rathnavel
Sent: Tuesday, December 6, 2016 8:49 PM
To: Storage Performance Development Kit <spdk(a)lists.01.org>
Subject: Re: [SPDK] Performance

The setup is the default setup of NVMf target with Direct attached Subsystem. The FIO plugin was at the target side and the IOPS from SPDK was found to be around 2.5K and the IOPS from kernel was found to be 100k.

Regards,
Param.
On 06-Dec-2016, at 5:57 PM, Andrey Kuzmin <andrey.v.kuzmin(a)gmail.com<mailto:andrey.v.kuzmin(a)gmail.com>> wrote:


On Tue, Dec 6, 2016, 15:24 Kumaraparameshwaran Rathnavel <krath(a)cloudsimple.com<mailto:krath(a)cloudsimple.com>> wrote:
Hi All,

Is there any benchmarking done to prove that SPDK is better than the Kernel driver?

We did run the sample driver provided from the SPDK and compared with the kernel interface and found that SPDK is very slow. We did run FIO.

It's worth publishing your setup and results first.

Regards,
Andrey

Can you describe the setup for the performance measurement between Kernel and SPDK driver.

Regards,
Param.
_______________________________________________
SPDK mailing list
SPDK(a)lists.01.org<mailto:SPDK(a)lists.01.org>
https://lists.01.org/mailman/listinfo/spdk
--
Regards,
Andrey
_______________________________________________
SPDK mailing list
SPDK(a)lists.01.org<mailto:SPDK(a)lists.01.org>
https://lists.01.org/mailman/listinfo/spdk


[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 5910 bytes --]

^ permalink raw reply	[flat|nested] 12+ messages in thread
* Re: [SPDK] Performance
@ 2016-12-06 16:35 Harris, James R
  0 siblings, 0 replies; 12+ messages in thread
From: Harris, James R @ 2016-12-06 16:35 UTC (permalink / raw)
  To: spdk

[-- Attachment #1: Type: text/plain, Size: 1850 bytes --]


On Dec 6, 2016, at 5:49 AM, Kumaraparameshwaran Rathnavel <krath(a)cloudsimple.com<mailto:krath(a)cloudsimple.com>> wrote:

The setup is the default setup of NVMf target with Direct attached Subsystem. The FIO plugin was at the target side and the IOPS from SPDK was found to be around 2.5K and the IOPS from kernel was found to be 100k.


Hi Param,

Can you check the AcceptorPollRate setting in your configuration file?  There was a recent fix to the default value for this parameter.  You should set it to at least 10000 (10ms).

https://github.com/spdk/spdk/commit/5f3761cf7ec50cf03183203dc50657a9f2a18908

Thanks,

-Jim


Regards,
Param.
On 06-Dec-2016, at 5:57 PM, Andrey Kuzmin <andrey.v.kuzmin(a)gmail.com<mailto:andrey.v.kuzmin(a)gmail.com>> wrote:



On Tue, Dec 6, 2016, 15:24 Kumaraparameshwaran Rathnavel <krath(a)cloudsimple.com<mailto:krath(a)cloudsimple.com>> wrote:
Hi All,

Is there any benchmarking done to prove that SPDK is better than the Kernel driver?

We did run the sample driver provided from the SPDK and compared with the kernel interface and found that SPDK is very slow. We did run FIO.

It's worth publishing your setup and results first.

Regards,
Andrey

Can you describe the setup for the performance measurement between Kernel and SPDK driver.

Regards,
Param.
_______________________________________________
SPDK mailing list
SPDK(a)lists.01.org<mailto:SPDK(a)lists.01.org>
https://lists.01.org/mailman/listinfo/spdk
--

Regards,
Andrey

_______________________________________________
SPDK mailing list
SPDK(a)lists.01.org<mailto:SPDK(a)lists.01.org>
https://lists.01.org/mailman/listinfo/spdk

_______________________________________________
SPDK mailing list
SPDK(a)lists.01.org<mailto:SPDK(a)lists.01.org>
https://lists.01.org/mailman/listinfo/spdk


[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 4435 bytes --]

^ permalink raw reply	[flat|nested] 12+ messages in thread
* Re: [SPDK] Performance
@ 2016-12-06 13:20 Andrey Kuzmin
  0 siblings, 0 replies; 12+ messages in thread
From: Andrey Kuzmin @ 2016-12-06 13:20 UTC (permalink / raw)
  To: spdk

[-- Attachment #1: Type: text/plain, Size: 3334 bytes --]

There hasn't been ay reactor stuff a year ago. The perf you claim seems so
way off what I saw that you're likely dealing with some major
misconfiguration issue.

On Dec 6, 2016 16:08, "Kumaraparameshwaran Rathnavel" <krath(a)cloudsimple.com>
wrote:

> What was the Reactor mask and did you assign different cores ? What was
> the queues per connection number in the target side? I had all as default
> values.
>
> Regards,
> Param.
>
> On 06-Dec-2016, at 6:35 PM, Andrey Kuzmin <andrey.v.kuzmin(a)gmail.com>
> wrote:
>
> Same as yours, essentially, although run about a year ago.
>
> On Dec 6, 2016 15:57, "Kumaraparameshwaran Rathnavel" <
> krath(a)cloudsimple.com> wrote:
>
>> Can you briefly describe your setup if you don’t mind.
>>
>> It will be of great help.
>>
>> Regards,
>> Param.
>>
>> On 06-Dec-2016, at 6:24 PM, Andrey Kuzmin <andrey.v.kuzmin(a)gmail.com>
>> wrote:
>>
>>
>>
>> On Tue, Dec 6, 2016, 15:49 Kumaraparameshwaran Rathnavel <
>> krath(a)cloudsimple.com> wrote:
>>
>>> The setup is the default setup of NVMf target with Direct attached
>>> Subsystem. The FIO plugin was at the target side and the IOPS from SPDK was
>>> found to be around 2.5K and the IOPS from kernel was found to be 100k.
>>>
>>
>> Sounds like a setup issue to me. In my tests, SPDK typically matched or
>> surpassed kernel stack in IOPS while using 2-3x less CPU.
>>
>> Regards,
>> Andrey
>>
>>>
>>> Regards,
>>> Param.
>>>
>>> On 06-Dec-2016, at 5:57 PM, Andrey Kuzmin <andrey.v.kuzmin(a)gmail.com>
>>> wrote:
>>>
>>>
>>>
>>> On Tue, Dec 6, 2016, 15:24 Kumaraparameshwaran Rathnavel <
>>> krath(a)cloudsimple.com> wrote:
>>>
>>> Hi All,
>>>
>>> Is there any benchmarking done to prove that SPDK is better than the
>>> Kernel driver?
>>>
>>> We did run the sample driver provided from the SPDK and compared with
>>> the kernel interface and found that SPDK is very slow. We did run FIO.
>>>
>>>
>>> It's worth publishing your setup and results first.
>>>
>>> Regards,
>>> Andrey
>>>
>>>
>>> Can you describe the setup for the performance measurement between
>>> Kernel and SPDK driver.
>>>
>>> Regards,
>>> Param.
>>> _______________________________________________
>>> SPDK mailing list
>>> SPDK(a)lists.01.org
>>> https://lists.01.org/mailman/listinfo/spdk
>>>
>>> --
>>>
>>> Regards,
>>> Andrey
>>> _______________________________________________
>>> SPDK mailing list
>>> SPDK(a)lists.01.org
>>> https://lists.01.org/mailman/listinfo/spdk
>>>
>>>
>>> _______________________________________________
>>> SPDK mailing list
>>> SPDK(a)lists.01.org
>>> https://lists.01.org/mailman/listinfo/spdk
>>>
>> --
>>
>> Regards,
>> Andrey
>> _______________________________________________
>> SPDK mailing list
>> SPDK(a)lists.01.org
>> https://lists.01.org/mailman/listinfo/spdk
>>
>>
>>
>> _______________________________________________
>> SPDK mailing list
>> SPDK(a)lists.01.org
>> https://lists.01.org/mailman/listinfo/spdk
>>
>> _______________________________________________
> SPDK mailing list
> SPDK(a)lists.01.org
> https://lists.01.org/mailman/listinfo/spdk
>
>
>
> _______________________________________________
> SPDK mailing list
> SPDK(a)lists.01.org
> https://lists.01.org/mailman/listinfo/spdk
>
>

[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 14918 bytes --]

^ permalink raw reply	[flat|nested] 12+ messages in thread
* Re: [SPDK] Performance
@ 2016-12-06 13:08 Kumaraparameshwaran Rathnavel
  0 siblings, 0 replies; 12+ messages in thread
From: Kumaraparameshwaran Rathnavel @ 2016-12-06 13:08 UTC (permalink / raw)
  To: spdk

[-- Attachment #1: Type: text/plain, Size: 3340 bytes --]

What was the Reactor mask and did you assign different cores\x10? What was the queues per connection number in the target side? I had all as default values.

Regards,
Param.
> On 06-Dec-2016, at 6:35 PM, Andrey Kuzmin <andrey.v.kuzmin(a)gmail.com> wrote:
> 
> Same as yours, essentially, although run about a year ago.
> 
> On Dec 6, 2016 15:57, "Kumaraparameshwaran Rathnavel" <krath(a)cloudsimple.com <mailto:krath(a)cloudsimple.com>> wrote:
> Can you briefly describe your setup if you don’t mind.
> 
> It will be of great help.
> 
> Regards,
> Param.
>> On 06-Dec-2016, at 6:24 PM, Andrey Kuzmin <andrey.v.kuzmin(a)gmail.com <mailto:andrey.v.kuzmin(a)gmail.com>> wrote:
>> 
>> 
>> 
>> On Tue, Dec 6, 2016, 15:49 Kumaraparameshwaran Rathnavel <krath(a)cloudsimple.com <mailto:krath(a)cloudsimple.com>> wrote:
>> The setup is the default setup of NVMf target with Direct attached Subsystem. The FIO plugin was at the target side and the IOPS from SPDK was found to be around 2.5K and the IOPS from kernel was found to be 100k.
>> 
>> Sounds like a setup issue to me. In my tests, SPDK typically matched or surpassed kernel stack in IOPS while using 2-3x less CPU.
>> 
>> Regards,
>> Andrey
>> 
>> Regards,
>> Param.
>> 
>>> On 06-Dec-2016, at 5:57 PM, Andrey Kuzmin <andrey.v.kuzmin(a)gmail.com <mailto:andrey.v.kuzmin(a)gmail.com>> wrote:
>>> 
>>> 
>>> 
>>> On Tue, Dec 6, 2016, 15:24 Kumaraparameshwaran Rathnavel <krath(a)cloudsimple.com <mailto:krath(a)cloudsimple.com>> wrote:
>>> Hi All,
>>> 
>>> Is there any benchmarking done to prove that SPDK is better than the Kernel driver?
>>> 
>>> We did run the sample driver provided from the SPDK and compared with the kernel interface and found that SPDK is very slow. We did run FIO.
>>> 
>>> It's worth publishing your setup and results first.
>>> 
>>> Regards,
>>> Andrey
>>> 
>>> Can you describe the setup for the performance measurement between Kernel and SPDK driver.
>>> 
>>> Regards,
>>> Param.
>>> _______________________________________________
>>> SPDK mailing list
>>> SPDK(a)lists.01.org <mailto:SPDK(a)lists.01.org>
>>> https://lists.01.org/mailman/listinfo/spdk <https://lists.01.org/mailman/listinfo/spdk>
>>> -- 
>>> Regards,
>>> Andrey
>>> 
>>> _______________________________________________
>>> SPDK mailing list
>>> SPDK(a)lists.01.org <mailto:SPDK(a)lists.01.org>
>>> https://lists.01.org/mailman/listinfo/spdk <https://lists.01.org/mailman/listinfo/spdk>
>> 
>> _______________________________________________
>> SPDK mailing list
>> SPDK(a)lists.01.org <mailto:SPDK(a)lists.01.org>
>> https://lists.01.org/mailman/listinfo/spdk <https://lists.01.org/mailman/listinfo/spdk>
>> -- 
>> Regards,
>> Andrey
>> 
>> _______________________________________________
>> SPDK mailing list
>> SPDK(a)lists.01.org <mailto:SPDK(a)lists.01.org>
>> https://lists.01.org/mailman/listinfo/spdk <https://lists.01.org/mailman/listinfo/spdk>
> 
> _______________________________________________
> SPDK mailing list
> SPDK(a)lists.01.org <mailto:SPDK(a)lists.01.org>
> https://lists.01.org/mailman/listinfo/spdk <https://lists.01.org/mailman/listinfo/spdk>
> 
> _______________________________________________
> SPDK mailing list
> SPDK(a)lists.01.org
> https://lists.01.org/mailman/listinfo/spdk


[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 13321 bytes --]

^ permalink raw reply	[flat|nested] 12+ messages in thread
* Re: [SPDK] Performance
@ 2016-12-06 13:05 Andrey Kuzmin
  0 siblings, 0 replies; 12+ messages in thread
From: Andrey Kuzmin @ 2016-12-06 13:05 UTC (permalink / raw)
  To: spdk

[-- Attachment #1: Type: text/plain, Size: 2400 bytes --]

Same as yours, essentially, although run about a year ago.

On Dec 6, 2016 15:57, "Kumaraparameshwaran Rathnavel" <krath(a)cloudsimple.com>
wrote:

> Can you briefly describe your setup if you don’t mind.
>
> It will be of great help.
>
> Regards,
> Param.
>
> On 06-Dec-2016, at 6:24 PM, Andrey Kuzmin <andrey.v.kuzmin(a)gmail.com>
> wrote:
>
>
>
> On Tue, Dec 6, 2016, 15:49 Kumaraparameshwaran Rathnavel <
> krath(a)cloudsimple.com> wrote:
>
>> The setup is the default setup of NVMf target with Direct attached
>> Subsystem. The FIO plugin was at the target side and the IOPS from SPDK was
>> found to be around 2.5K and the IOPS from kernel was found to be 100k.
>>
>
> Sounds like a setup issue to me. In my tests, SPDK typically matched or
> surpassed kernel stack in IOPS while using 2-3x less CPU.
>
> Regards,
> Andrey
>
>>
>> Regards,
>> Param.
>>
>> On 06-Dec-2016, at 5:57 PM, Andrey Kuzmin <andrey.v.kuzmin(a)gmail.com>
>> wrote:
>>
>>
>>
>> On Tue, Dec 6, 2016, 15:24 Kumaraparameshwaran Rathnavel <
>> krath(a)cloudsimple.com> wrote:
>>
>> Hi All,
>>
>> Is there any benchmarking done to prove that SPDK is better than the
>> Kernel driver?
>>
>> We did run the sample driver provided from the SPDK and compared with the
>> kernel interface and found that SPDK is very slow. We did run FIO.
>>
>>
>> It's worth publishing your setup and results first.
>>
>> Regards,
>> Andrey
>>
>>
>> Can you describe the setup for the performance measurement between Kernel
>> and SPDK driver.
>>
>> Regards,
>> Param.
>> _______________________________________________
>> SPDK mailing list
>> SPDK(a)lists.01.org
>> https://lists.01.org/mailman/listinfo/spdk
>>
>> --
>>
>> Regards,
>> Andrey
>> _______________________________________________
>> SPDK mailing list
>> SPDK(a)lists.01.org
>> https://lists.01.org/mailman/listinfo/spdk
>>
>>
>> _______________________________________________
>> SPDK mailing list
>> SPDK(a)lists.01.org
>> https://lists.01.org/mailman/listinfo/spdk
>>
> --
>
> Regards,
> Andrey
> _______________________________________________
> SPDK mailing list
> SPDK(a)lists.01.org
> https://lists.01.org/mailman/listinfo/spdk
>
>
>
> _______________________________________________
> SPDK mailing list
> SPDK(a)lists.01.org
> https://lists.01.org/mailman/listinfo/spdk
>
>

[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 11806 bytes --]

^ permalink raw reply	[flat|nested] 12+ messages in thread
* Re: [SPDK] Performance
@ 2016-12-06 12:57 Kumaraparameshwaran Rathnavel
  0 siblings, 0 replies; 12+ messages in thread
From: Kumaraparameshwaran Rathnavel @ 2016-12-06 12:57 UTC (permalink / raw)
  To: spdk

[-- Attachment #1: Type: text/plain, Size: 2415 bytes --]

Can you briefly describe your setup if you don’t mind.

It will be of great help.

Regards,
Param.
> On 06-Dec-2016, at 6:24 PM, Andrey Kuzmin <andrey.v.kuzmin(a)gmail.com> wrote:
> 
> 
> 
> On Tue, Dec 6, 2016, 15:49 Kumaraparameshwaran Rathnavel <krath(a)cloudsimple.com <mailto:krath(a)cloudsimple.com>> wrote:
> The setup is the default setup of NVMf target with Direct attached Subsystem. The FIO plugin was at the target side and the IOPS from SPDK was found to be around 2.5K and the IOPS from kernel was found to be 100k.
> 
> Sounds like a setup issue to me. In my tests, SPDK typically matched or surpassed kernel stack in IOPS while using 2-3x less CPU.
> 
> Regards,
> Andrey
> 
> Regards,
> Param.
> 
>> On 06-Dec-2016, at 5:57 PM, Andrey Kuzmin <andrey.v.kuzmin(a)gmail.com <mailto:andrey.v.kuzmin(a)gmail.com>> wrote:
>> 
>> 
>> 
>> On Tue, Dec 6, 2016, 15:24 Kumaraparameshwaran Rathnavel <krath(a)cloudsimple.com <mailto:krath(a)cloudsimple.com>> wrote:
>> Hi All,
>> 
>> Is there any benchmarking done to prove that SPDK is better than the Kernel driver?
>> 
>> We did run the sample driver provided from the SPDK and compared with the kernel interface and found that SPDK is very slow. We did run FIO.
>> 
>> It's worth publishing your setup and results first.
>> 
>> Regards,
>> Andrey
>> 
>> Can you describe the setup for the performance measurement between Kernel and SPDK driver.
>> 
>> Regards,
>> Param.
>> _______________________________________________
>> SPDK mailing list
>> SPDK(a)lists.01.org <mailto:SPDK(a)lists.01.org>
>> https://lists.01.org/mailman/listinfo/spdk <https://lists.01.org/mailman/listinfo/spdk>
>> -- 
>> Regards,
>> Andrey
>> 
>> _______________________________________________
>> SPDK mailing list
>> SPDK(a)lists.01.org <mailto:SPDK(a)lists.01.org>
>> https://lists.01.org/mailman/listinfo/spdk <https://lists.01.org/mailman/listinfo/spdk>
> 
> _______________________________________________
> SPDK mailing list
> SPDK(a)lists.01.org <mailto:SPDK(a)lists.01.org>
> https://lists.01.org/mailman/listinfo/spdk <https://lists.01.org/mailman/listinfo/spdk>
> -- 
> Regards,
> Andrey
> 
> _______________________________________________
> SPDK mailing list
> SPDK(a)lists.01.org <mailto:SPDK(a)lists.01.org>
> https://lists.01.org/mailman/listinfo/spdk <https://lists.01.org/mailman/listinfo/spdk>

[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 11664 bytes --]

^ permalink raw reply	[flat|nested] 12+ messages in thread
* Re: [SPDK] Performance
@ 2016-12-06 12:54 Andrey Kuzmin
  0 siblings, 0 replies; 12+ messages in thread
From: Andrey Kuzmin @ 2016-12-06 12:54 UTC (permalink / raw)
  To: spdk

[-- Attachment #1: Type: text/plain, Size: 1658 bytes --]

On Tue, Dec 6, 2016, 15:49 Kumaraparameshwaran Rathnavel <
krath(a)cloudsimple.com> wrote:

> The setup is the default setup of NVMf target with Direct attached
> Subsystem. The FIO plugin was at the target side and the IOPS from SPDK was
> found to be around 2.5K and the IOPS from kernel was found to be 100k.
>

Sounds like a setup issue to me. In my tests, SPDK typically matched or
surpassed kernel stack in IOPS while using 2-3x less CPU.

Regards,
Andrey

>
> Regards,
> Param.
>
> On 06-Dec-2016, at 5:57 PM, Andrey Kuzmin <andrey.v.kuzmin(a)gmail.com>
> wrote:
>
>
>
> On Tue, Dec 6, 2016, 15:24 Kumaraparameshwaran Rathnavel <
> krath(a)cloudsimple.com> wrote:
>
> Hi All,
>
> Is there any benchmarking done to prove that SPDK is better than the
> Kernel driver?
>
> We did run the sample driver provided from the SPDK and compared with the
> kernel interface and found that SPDK is very slow. We did run FIO.
>
>
> It's worth publishing your setup and results first.
>
> Regards,
> Andrey
>
>
> Can you describe the setup for the performance measurement between Kernel
> and SPDK driver.
>
> Regards,
> Param.
> _______________________________________________
> SPDK mailing list
> SPDK(a)lists.01.org
> https://lists.01.org/mailman/listinfo/spdk
>
> --
>
> Regards,
> Andrey
> _______________________________________________
> SPDK mailing list
> SPDK(a)lists.01.org
> https://lists.01.org/mailman/listinfo/spdk
>
>
> _______________________________________________
> SPDK mailing list
> SPDK(a)lists.01.org
> https://lists.01.org/mailman/listinfo/spdk
>
-- 

Regards,
Andrey

[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 4531 bytes --]

^ permalink raw reply	[flat|nested] 12+ messages in thread
* Re: [SPDK] Performance
@ 2016-12-06 12:49 Kumaraparameshwaran Rathnavel
  0 siblings, 0 replies; 12+ messages in thread
From: Kumaraparameshwaran Rathnavel @ 2016-12-06 12:49 UTC (permalink / raw)
  To: spdk

[-- Attachment #1: Type: text/plain, Size: 1305 bytes --]

The setup is the default setup of NVMf target with Direct attached Subsystem. The FIO plugin was at the target side and the IOPS from SPDK was found to be around 2.5K and the IOPS from kernel was found to be 100k.

Regards,
Param.
> On 06-Dec-2016, at 5:57 PM, Andrey Kuzmin <andrey.v.kuzmin(a)gmail.com> wrote:
> 
> 
> 
> On Tue, Dec 6, 2016, 15:24 Kumaraparameshwaran Rathnavel <krath(a)cloudsimple.com <mailto:krath(a)cloudsimple.com>> wrote:
> Hi All,
> 
> Is there any benchmarking done to prove that SPDK is better than the Kernel driver?
> 
> We did run the sample driver provided from the SPDK and compared with the kernel interface and found that SPDK is very slow. We did run FIO.
> 
> It's worth publishing your setup and results first.
> 
> Regards,
> Andrey
> 
> Can you describe the setup for the performance measurement between Kernel and SPDK driver.
> 
> Regards,
> Param.
> _______________________________________________
> SPDK mailing list
> SPDK(a)lists.01.org <mailto:SPDK(a)lists.01.org>
> https://lists.01.org/mailman/listinfo/spdk <https://lists.01.org/mailman/listinfo/spdk>
> -- 
> Regards,
> Andrey
> 
> _______________________________________________
> SPDK mailing list
> SPDK(a)lists.01.org
> https://lists.01.org/mailman/listinfo/spdk


[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 2847 bytes --]

^ permalink raw reply	[flat|nested] 12+ messages in thread
* Re: [SPDK] Performance
@ 2016-12-06 12:27 Andrey Kuzmin
  0 siblings, 0 replies; 12+ messages in thread
From: Andrey Kuzmin @ 2016-12-06 12:27 UTC (permalink / raw)
  To: spdk

[-- Attachment #1: Type: text/plain, Size: 723 bytes --]

On Tue, Dec 6, 2016, 15:24 Kumaraparameshwaran Rathnavel <
krath(a)cloudsimple.com> wrote:

> Hi All,
>
> Is there any benchmarking done to prove that SPDK is better than the
> Kernel driver?
>
> We did run the sample driver provided from the SPDK and compared with the
> kernel interface and found that SPDK is very slow. We did run FIO.
>

It's worth publishing your setup and results first.

Regards,
Andrey

>
> Can you describe the setup for the performance measurement between Kernel
> and SPDK driver.
>
> Regards,
> Param.
> _______________________________________________
> SPDK mailing list
> SPDK(a)lists.01.org
> https://lists.01.org/mailman/listinfo/spdk
>
-- 

Regards,
Andrey

[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 1663 bytes --]

^ permalink raw reply	[flat|nested] 12+ messages in thread
* [SPDK] Performance
@ 2016-12-06 12:24 Kumaraparameshwaran Rathnavel
  0 siblings, 0 replies; 12+ messages in thread
From: Kumaraparameshwaran Rathnavel @ 2016-12-06 12:24 UTC (permalink / raw)
  To: spdk

[-- Attachment #1: Type: text/plain, Size: 353 bytes --]

Hi All,

Is there any benchmarking done to prove that SPDK is better than the Kernel driver? 

We did run the sample driver provided from the SPDK and compared with the kernel interface and found that SPDK is very slow. We did run FIO.

Can you describe the setup for the performance measurement between Kernel and SPDK driver.

Regards,
Param.

^ permalink raw reply	[flat|nested] 12+ messages in thread

end of thread, other threads:[~2017-01-18 23:41 UTC | newest]

Thread overview: 12+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-12-06 16:58 [SPDK] Performance Walker, Benjamin
  -- strict thread matches above, loose matches on Subject: below --
2017-01-18 23:41 Walker, Benjamin
2016-12-07  1:22 Yang, Ziye
2016-12-06 16:35 Harris, James R
2016-12-06 13:20 Andrey Kuzmin
2016-12-06 13:08 Kumaraparameshwaran Rathnavel
2016-12-06 13:05 Andrey Kuzmin
2016-12-06 12:57 Kumaraparameshwaran Rathnavel
2016-12-06 12:54 Andrey Kuzmin
2016-12-06 12:49 Kumaraparameshwaran Rathnavel
2016-12-06 12:27 Andrey Kuzmin
2016-12-06 12:24 Kumaraparameshwaran Rathnavel

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.