Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-18 Thread Kent Overstreet
On Tue, Dec 18, 2012 at 11:16 AM, Kent Overstreet wrote: > Or maybe just getting rid of the ringbuffer is that awesome. Gonna try > and work on combining our optimizations so I can see what that looks > like :) Yes, yes it is. Combined our aio/dio patches and got 50% better throughput than I'd

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-18 Thread Kent Overstreet
On Sat, Dec 15, 2012 at 02:16:48PM +0100, Jens Axboe wrote: > On 2012-12-15 11:36, Kent Overstreet wrote: > >> Knock yourself out - I already took a quick look at it, and conversion > >> should be pretty simple. It's the mtip32xx driver, it's in the kernel. I > >> would suggest getting rid of the

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-18 Thread Kent Overstreet
On Sat, Dec 15, 2012 at 02:16:48PM +0100, Jens Axboe wrote: On 2012-12-15 11:36, Kent Overstreet wrote: Knock yourself out - I already took a quick look at it, and conversion should be pretty simple. It's the mtip32xx driver, it's in the kernel. I would suggest getting rid of the

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-18 Thread Kent Overstreet
On Tue, Dec 18, 2012 at 11:16 AM, Kent Overstreet koverstr...@google.com wrote: Or maybe just getting rid of the ringbuffer is that awesome. Gonna try and work on combining our optimizations so I can see what that looks like :) Yes, yes it is. Combined our aio/dio patches and got 50% better

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-15 Thread Jens Axboe
On 2012-12-15 11:36, Kent Overstreet wrote: >> Knock yourself out - I already took a quick look at it, and conversion >> should be pretty simple. It's the mtip32xx driver, it's in the kernel. I >> would suggest getting rid of the ->async_callback() (since it's always >> bio_endio()) since that'll

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-15 Thread Jens Axboe
On 2012-12-15 11:36, Kent Overstreet wrote: > On Sat, Dec 15, 2012 at 10:46:32AM +0100, Jens Axboe wrote: >> On 2012-12-15 10:25, Kent Overstreet wrote: >>> Cool, thanks for the numbers! >>> >>> I suspect the difference is due to contention on the ringbuffer, >>> completion side. You didn't enable

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-15 Thread Kent Overstreet
On Sat, Dec 15, 2012 at 10:46:32AM +0100, Jens Axboe wrote: > On 2012-12-15 10:25, Kent Overstreet wrote: > > Cool, thanks for the numbers! > > > > I suspect the difference is due to contention on the ringbuffer, > > completion side. You didn't enable my batched completion stuff, did you? > >

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-15 Thread Jens Axboe
On 2012-12-15 10:25, Kent Overstreet wrote: > On Fri, Dec 14, 2012 at 08:35:53AM +0100, Jens Axboe wrote: >> On 2012-12-14 03:26, Jack Wang wrote: >>> 2012/12/14 Jens Axboe : On Mon, Dec 03 2012, Kent Overstreet wrote: > Last posting:

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-15 Thread Kent Overstreet
On Fri, Dec 14, 2012 at 08:35:53AM +0100, Jens Axboe wrote: > On 2012-12-14 03:26, Jack Wang wrote: > > 2012/12/14 Jens Axboe : > >> On Mon, Dec 03 2012, Kent Overstreet wrote: > >>> Last posting: http://thread.gmane.org/gmane.linux.kernel.aio.general/3169 > >>> > >>> Changes since the last

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-15 Thread Kent Overstreet
On Fri, Dec 14, 2012 at 08:35:53AM +0100, Jens Axboe wrote: On 2012-12-14 03:26, Jack Wang wrote: 2012/12/14 Jens Axboe jax...@fusionio.com: On Mon, Dec 03 2012, Kent Overstreet wrote: Last posting: http://thread.gmane.org/gmane.linux.kernel.aio.general/3169 Changes since the last

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-15 Thread Jens Axboe
On 2012-12-15 10:25, Kent Overstreet wrote: On Fri, Dec 14, 2012 at 08:35:53AM +0100, Jens Axboe wrote: On 2012-12-14 03:26, Jack Wang wrote: 2012/12/14 Jens Axboe jax...@fusionio.com: On Mon, Dec 03 2012, Kent Overstreet wrote: Last posting:

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-15 Thread Kent Overstreet
On Sat, Dec 15, 2012 at 10:46:32AM +0100, Jens Axboe wrote: On 2012-12-15 10:25, Kent Overstreet wrote: Cool, thanks for the numbers! I suspect the difference is due to contention on the ringbuffer, completion side. You didn't enable my batched completion stuff, did you? No, haven't

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-15 Thread Jens Axboe
On 2012-12-15 11:36, Kent Overstreet wrote: On Sat, Dec 15, 2012 at 10:46:32AM +0100, Jens Axboe wrote: On 2012-12-15 10:25, Kent Overstreet wrote: Cool, thanks for the numbers! I suspect the difference is due to contention on the ringbuffer, completion side. You didn't enable my batched

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-15 Thread Jens Axboe
On 2012-12-15 11:36, Kent Overstreet wrote: Knock yourself out - I already took a quick look at it, and conversion should be pretty simple. It's the mtip32xx driver, it's in the kernel. I would suggest getting rid of the -async_callback() (since it's always bio_endio()) since that'll make it

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-13 Thread Jens Axboe
On 2012-12-14 03:26, Jack Wang wrote: > 2012/12/14 Jens Axboe : >> On Mon, Dec 03 2012, Kent Overstreet wrote: >>> Last posting: http://thread.gmane.org/gmane.linux.kernel.aio.general/3169 >>> >>> Changes since the last posting should all be noted in the individual >>> patch descriptions. >>> >>>

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-13 Thread Jack Wang
2012/12/14 Jens Axboe : > On Mon, Dec 03 2012, Kent Overstreet wrote: >> Last posting: http://thread.gmane.org/gmane.linux.kernel.aio.general/3169 >> >> Changes since the last posting should all be noted in the individual >> patch descriptions. >> >> * Zach pointed out the aio_read_evt() patch

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-13 Thread Jens Axboe
On Mon, Dec 03 2012, Kent Overstreet wrote: > Last posting: http://thread.gmane.org/gmane.linux.kernel.aio.general/3169 > > Changes since the last posting should all be noted in the individual > patch descriptions. > > * Zach pointed out the aio_read_evt() patch was calling functions that >

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-13 Thread Jens Axboe
On Mon, Dec 03 2012, Kent Overstreet wrote: Last posting: http://thread.gmane.org/gmane.linux.kernel.aio.general/3169 Changes since the last posting should all be noted in the individual patch descriptions. * Zach pointed out the aio_read_evt() patch was calling functions that could

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-13 Thread Jack Wang
2012/12/14 Jens Axboe jax...@fusionio.com: On Mon, Dec 03 2012, Kent Overstreet wrote: Last posting: http://thread.gmane.org/gmane.linux.kernel.aio.general/3169 Changes since the last posting should all be noted in the individual patch descriptions. * Zach pointed out the aio_read_evt()

Re: [PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-13 Thread Jens Axboe
On 2012-12-14 03:26, Jack Wang wrote: 2012/12/14 Jens Axboe jax...@fusionio.com: On Mon, Dec 03 2012, Kent Overstreet wrote: Last posting: http://thread.gmane.org/gmane.linux.kernel.aio.general/3169 Changes since the last posting should all be noted in the individual patch descriptions. *

[PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-03 Thread Kent Overstreet
Last posting: http://thread.gmane.org/gmane.linux.kernel.aio.general/3169 Changes since the last posting should all be noted in the individual patch descriptions. * Zach pointed out the aio_read_evt() patch was calling functions that could sleep in TASK_INTERRUPTIBLE state, that patch is

[PATCH 00/26] AIO performance improvements/cleanups, v2

2012-12-03 Thread Kent Overstreet
Last posting: http://thread.gmane.org/gmane.linux.kernel.aio.general/3169 Changes since the last posting should all be noted in the individual patch descriptions. * Zach pointed out the aio_read_evt() patch was calling functions that could sleep in TASK_INTERRUPTIBLE state, that patch is