Re: [MLton] follow-up [cost of _export]

On 12/17/14 17:10, Matthew Fluet wrote:
> On Tue, Dec 9, 2014 at 10:59 AM, Bernard Berthomieu <[email protected]> wrote:
>> The total time drops to 8s with "fixed-heap 1G", but rises to
>> 500s (99% of which in gc) with "copy-generational-ratio 8.0"
>> (default for my applications) with a huge amount of minor gc.
> I'm not exactly sure why the generational GC is so problematic.
> Possibly due to the fact that all of the created threads almost
> certainly need to have their stack resized and a resized stack is
> pre-tenured in the old generation.  Thus, the minor GCs miss much of
> the actual garbage.

I re-run the example with gc-messages, gc-summary and
copy-generational-ratio 8.0.

gc-summary tells:

GC type        time ms     number          bytes          bytes/sec
-------------    -------    -------    --------------- ---------------
copying         14,218        759      2,034,262,364 143,076,550
mark-compact          0          0                  0               -
minor        419,078    855,134        363,163,964 866,578
total time: 693,004 ms
total GC time: 521,129 ms (75.2%)
max pause time: 138 ms
total bytes allocated: 14,274,555,396 bytes
max bytes live: 14,033,600 bytes
max heap size: 111,583,232 bytes
max stack size: 9,000 bytes
num cards marked: 6,694,859
bytes scanned: 39,805,104,216 bytes
bytes hash consed: 0 bytes


In case you want to dig deeper, I put the compressed log file
with gc-messages at:

     www.laas.fr/tina/tmp/ARM.log.gz

(It is 1,5 Gb uncompressed).

Best,
   Bernard.


------------------------------------------------------------------------------
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk