[PATCH v3 00/15] mm/memory: optimize fork() with PTE-mapped THP
Ryan Roberts
ryan.roberts at arm.com
Thu Feb 1 00:16:48 AEDT 2024
On 31/01/2024 12:56, David Hildenbrand wrote:
> On 31.01.24 13:37, Ryan Roberts wrote:
>> On 31/01/2024 11:49, Ryan Roberts wrote:
>>> On 31/01/2024 11:28, David Hildenbrand wrote:
>>>> On 31.01.24 12:16, Ryan Roberts wrote:
>>>>> On 31/01/2024 11:06, David Hildenbrand wrote:
>>>>>> On 31.01.24 11:43, Ryan Roberts wrote:
>>>>>>> On 29/01/2024 12:46, David Hildenbrand wrote:
>>>>>>>> Now that the rmap overhaul[1] is upstream that provides a clean interface
>>>>>>>> for rmap batching, let's implement PTE batching during fork when processing
>>>>>>>> PTE-mapped THPs.
>>>>>>>>
>>>>>>>> This series is partially based on Ryan's previous work[2] to implement
>>>>>>>> cont-pte support on arm64, but its a complete rewrite based on [1] to
>>>>>>>> optimize all architectures independent of any such PTE bits, and to
>>>>>>>> use the new rmap batching functions that simplify the code and prepare
>>>>>>>> for further rmap accounting changes.
>>>>>>>>
>>>>>>>> We collect consecutive PTEs that map consecutive pages of the same large
>>>>>>>> folio, making sure that the other PTE bits are compatible, and (a) adjust
>>>>>>>> the refcount only once per batch, (b) call rmap handling functions only
>>>>>>>> once per batch and (c) perform batch PTE setting/updates.
>>>>>>>>
>>>>>>>> While this series should be beneficial for adding cont-pte support on
>>>>>>>> ARM64[2], it's one of the requirements for maintaining a total mapcount[3]
>>>>>>>> for large folios with minimal added overhead and further changes[4] that
>>>>>>>> build up on top of the total mapcount.
>>>>>>>>
>>>>>>>> Independent of all that, this series results in a speedup during fork with
>>>>>>>> PTE-mapped THP, which is the default with THPs that are smaller than a PMD
>>>>>>>> (for example, 16KiB to 1024KiB mTHPs for anonymous memory[5]).
>>>>>>>>
>>>>>>>> On an Intel Xeon Silver 4210R CPU, fork'ing with 1GiB of PTE-mapped folios
>>>>>>>> of the same size (stddev < 1%) results in the following runtimes
>>>>>>>> for fork() (shorter is better):
>>>>>>>>
>>>>>>>> Folio Size | v6.8-rc1 | New | Change
>>>>>>>> ------------------------------------------
>>>>>>>> 4KiB | 0.014328 | 0.014035 | - 2%
>>>>>>>> 16KiB | 0.014263 | 0.01196 | -16%
>>>>>>>> 32KiB | 0.014334 | 0.01094 | -24%
>>>>>>>> 64KiB | 0.014046 | 0.010444 | -26%
>>>>>>>> 128KiB | 0.014011 | 0.010063 | -28%
>>>>>>>> 256KiB | 0.013993 | 0.009938 | -29%
>>>>>>>> 512KiB | 0.013983 | 0.00985 | -30%
>>>>>>>> 1024KiB | 0.013986 | 0.00982 | -30%
>>>>>>>> 2048KiB | 0.014305 | 0.010076 | -30%
>>>>>>>
>>>>>>> Just a heads up that I'm seeing some strange results on Apple M2. Fork for
>>>>>>> order-0 is seemingly costing ~17% more. I'm using GCC 13.2 and was pretty
>>>>>>> sure I
>>>>>>> didn't see this problem with version 1; although that was on a different
>>>>>>> baseline and I've thrown the numbers away so will rerun and try to debug
>>>>>>> this.
>>>
>>> Numbers for v1 of the series, both on top of 6.8-rc1 and rebased to the same
>>> mm-unstable base as v3 of the series (first 2 rows are from what I just posted
>>> for context):
>>>
>>> | kernel | mean_rel | std_rel |
>>> |:-------------------|-----------:|----------:|
>>> | mm-unstabe (base) | 0.0% | 1.1% |
>>> | mm-unstable + v3 | 16.7% | 0.8% |
>>> | mm-unstable + v1 | -2.5% | 1.7% |
>>> | v6.8-rc1 + v1 | -6.6% | 1.1% |
>>>
>>> So all looks good with v1. And seems to suggest mm-unstable has regressed by ~4%
>>> vs v6.8-rc1. Is this really a useful benchmark? Does the raw performance of
>>> fork() syscall really matter? Evidence suggests its moving all over the place -
>>> breath on the code and it changes - not a great place to be when using the test
>>> for gating purposes!
>>>
>>> Still with the old tests - I'll move to the new ones now.
>>>
>>>
>>>>>>>
>>>>>>
>>>>>> So far, on my x86 tests (Intel, AMD EPYC), I was not able to observe this.
>>>>>> fork() for order-0 was consistently effectively unchanged. Do you observe
>>>>>> that
>>>>>> on other ARM systems as well?
>>>>>
>>>>> Nope; running the exact same kernel binary and user space on Altra, I see
>>>>> sensible numbers;
>>>>>
>>>>> fork order-0: -1.3%
>>>>> fork order-9: -7.6%
>>>>> dontneed order-0: -0.5%
>>>>> dontneed order-9: 0.1%
>>>>> munmap order-0: 0.0%
>>>>> munmap order-9: -67.9%
>>>>>
>>>>> So I guess some pipelining issue that causes the M2 to stall more?
>>>>
>>>> With one effective added folio_test_large(), it could only be a code layout
>>>> problem? Or the compiler does something stupid, but you say that you run the
>>>> exact same kernel binary, so that doesn't make sense.
>>>
>>> Yup, same binary. We know this code is very sensitive - 1 cycle makes a big
>>> difference. So could easily be code layout, branch prediction, etc...
>>>
>>>>
>>>> I'm also surprised about the dontneed vs. munmap numbers.
>>>
>>> You mean the ones for Altra that I posted? (I didn't post any for M2). The altra
>>> numbers look ok to me; dontneed has no change, and munmap has no change for
>>> order-0 and is massively improved for order-9.
>>>
>>> Doesn't make any sense
>>>> (again, there was this VMA merging problem but it would still allow for
>>>> batching
>>>> within a single VMA that spans exactly one large folio).
>>>>
>>>> What are you using as baseline? Really just mm-unstable vs.
>>>> mm-unstable+patches?
>>>
>>> yes. except for "v6.8-rc1 + v1" above.
>>>
>>>>
>>>> Let's see if the new test changes the numbers you measure.
>>
>> Nope: looks the same. I've taken my test harness out of the picture and done
>> everything manually from the ground up, with the old tests and the new. Headline
>> is that I see similar numbers from both.
>
> I took me a while to get really reproducible numbers on Intel. Most importantly:
> * Set a fixed CPU frequency, disabling any boost and avoiding any
> thermal throttling.
> * Pin the test to CPUs and set a nice level.
I'm already pinning the test to cpu 0. But for M2, at least, I'm running in a VM
on top of macos, and I don't have a mechanism to pin the QEMU threads to the
physical CPUs. Anyway, I don't think these are problems because for a given
kernel build I can accurately repro numbers.
>
> Another thing is, to avoid systems where you can have NUMA effects within a
> single socket. Otherwise, memory access latency is just random and depends on
> what the buddy enjoys giving you.
Yep; same. M2 is 1 NUMA node. On Altra, I'm disabling the second NUMA node to
remove those effects.
>
> But you seem to get the same +17 even after reboots, so that indicates that the
> CPU is not happy about the code for some reason. And the weird thing is, that
> nothing significantly changed for order-0 folios between v1 and v3 that could
> explain any of this.
>
> I'm not worried about 5% or so, nobody cares. But it would be good to have at
> least an explanation why only that system shows +17%.
Yep understood.
>
>>
>> Some details:
>> - I'm running for 10 seconds then averaging the output
>
> Same here.
>
>> - test is bimodal; first run (of 10 seconds) after boot is a bit faster on
>> average (up to 10%) than the rest; I could guess this is due to the memory
>> being allocated more contiguously the first few times through, so struct
>> pages have better locality, but that's a guess.
>
> I think it also has to do with the PCP lists, and the high-pcp auto tuning (I
> played with disabling that). Running on a freshly booted system gave me
> reproducible results.
>
> But yes: I was observing something similar on AMD EPYC, where you get
> consecutive pages from the buddy, but once you allocate from the PCP it might no
> longer be consecutive.
>
>> - test is 5-10% slower when output is printed to terminal vs when redirected to
>> file. I've always effectively been redirecting. Not sure if this overhead
>> could start to dominate the regression and that's why you don't see it?
>
> That's weird, because we don't print while measuring? Anyhow, 5/10% variance on
> some system is not the end of the world.
I imagine its cache effects? More work to do to print the output could be
evicting some code that's in the benchmark path?
>
>>
>> I'm inclined to run this test for the last N kernel releases and if the number
>> moves around significantly, conclude that these tests don't really matter.
>> Otherwise its an exercise in randomly refactoring code until it works well, but
>> that's just overfitting to the compiler and hw. What do you think?
>
> Personally, I wouldn't lose sleep if you see weird, unexplainable behavior on
> some system (not even architecture!). Trying to optimize for that would indeed
> be random refactorings.
>
> But I would not be so fast to say that "these tests don't really matter" and
> then go wild and degrade them as much as you want. There are use cases that care
> about fork performance especially with order-0 pages -- such as Redis.
Indeed. But also remember that my fork baseline time is ~2.5ms, and I think you
said yours was 14ms :)
I'll continue to mess around with it until the end of the day. But I'm not
making any headway, then I'll change tack; I'll just measure the performance of
my contpte changes using your fork/zap stuff as the baseline and post based on that.
More information about the Linuxppc-dev
mailing list