VM(/cache) got worse from 2.4.4 -> 2.4.8

Matthias Schniedermeyer (ms@citd.de)
Sat, 11 Aug 2001 17:24:09 +0200 (MEST)


#Include <hallo.h>

System:
MB: Severworks HE-SL-Chipset
RAM: 3GB
CPU: 2xPIII 933Mhz
Relevant HDDs:
2x IBM DTLA 307045 connected via PDC20268(I patched the kernel with latest
IDE-driver i could find in people/hedrick)

Swap: NONE

When i do "big" (DATA>RAM) copies from HDD1 to HDD2 the system gets
"inresponsive" (Hitting keys are noticeable delayed (I'd says around
500ms)) when cache reaches maximum.

# free
total used free shared buffers cached
Mem: 3090664 3085452 5212 0 15156 2959200
-/+ buffers/cache: 111096 2979568
Swap: 0 0 0

The Sound(SB-Live, without the other patches for 2.4.8) hangs with the
EXACT same pattern as the blinking of the HDD-LED from where the data is
read. (LED goes out, sound hangs. LED lid sound plays)

Also the "Pattern" of cache "full" is diffrent from 2.4.4 to 2.4.8 when
2.4.4 reaches "maximun" then it "stops" and some MB of dirty caches are
written to target driver. (This could take severeal seconds) Whereas 2.4.8
uses very short "burstes" of reading source/writing target. (About 2/3 per
Second)

After the trashing situation is gone (Copy done) sound hanges exactly
every second. (Could be used as a timer. And doesn't get better until
reset)

2.4.4 didn't have ANY of those problems. (I haven't tried any kernel in
between)

OK. Time to downgrade to 2.4.4

Bis denn

-- 
Real Programmers consider "what you see is what you get" to be just as 
bad a concept in Text Editors as it is in women. No, the Real Programmer
wants a "you asked for it, you got it" text editor -- complicated, 
cryptic, powerful, unforgiving, dangerous.

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/