2.4 cpu usage...

I decided put 2.4 on my laptop. After getting config issues seemingly sorted out, still have some things I can't explain. VMware seems to run about 30% slower. X was even sluggish at times. When I'm doing 'nothing', top shows about 67% IDLE and 30% in 'system time'. I notice that the process "kapm-idled" is being counted as receiving alot of CPU time. Now this could make some sense maybe that idled is getting 30% of the time, but then there's the remaining 67% that came up idle.

I shut down X -- then top showed 5% idle and 95% in "kapm-idled" (and 95% system time) which could still make sense but is probably not the output you want to see when your computer is really idle.

So the kapm thing could be a "display" / accounting problem, but the slowdown in vmware/X was real. I ran a WIN Norton "Benchmark" -- comes up reliably over "300" -- usually around 320-350 under 2.2.17. Under 2.4, it came up reliably *under* 300 with typical being about 265".

So...I'm bummed. I'm assuming a 30% degradation in an app is probably not expected behavior? Swap usage is '0' in both OS's (i.e. it's not a run out of memory issue).

-l

--
L A Walsh | Trust Technology, Core Linux, SGI
[email protected] | Voice/Vmail: (650) 933-5338

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
Please read the FAQ at http://www.tux.org/lkml/

LA Walsh wrote:
[snip]

I actually saw a similar type of slowdown on my Inspiron 5000e with 2.4 (test10preX or test11preX, where I forget what X was, probably 5 or so).

Specifically, the following command read off the disk more slowly:

gzip -1 < /dev/hda | nc some_other_box some_port -w 1

even when I got rid of the gzip -1 (redirecting /dev/hda straight into nc, or piping cat into nc).

In my case, the slowdown was clearly visible - the disk light was at full brightness under 2.2.16, but not 2.4.0testwhatever.

I wanted to try again without any APM stuff in the kernel (or at least, not APM CPU idling) before reporting it, but I never got a chance to do that.

-Barry K. Nathan <[email protected]>
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
Please read the FAQ at http://www.tux.org/lkml/

I think we're on to something. I did gen's of the kernel with the following configs: 1) 2.2.17 (w/apm), 2) 2.4(w/apm), 3) 2.4(w/o apm). The apm seems to be a red herring in terms of actual performance hit. It seems to count apm time as 'system' time instead of 'idle'.

Kernel gen times were:

1) real    6m33.046s, user    6m1.380s, sys     0m18.450s
2)         6m45s real         6m12s user        0m26s sys  
3) real    6m49.275s, user    6m14.740s, sys    0m26.690s

The REAL problem was in disk performance. The apm made no difference:

hdparm -t /dev/hda1 /dev/hda3 /dev/hda4 /dev/hda5 /dev/hda7
1) 2.2.17

/dev/hda1: Timing buffered disk reads:  64 MB in  4.76 seconds = 13.45 MB/sec
/dev/hda3: Timing buffered disk reads:  64 MB in  4.57 seconds = 14.00 MB/sec
/dev/hda4: Timing buffered disk reads:  64 MB in  6.47 seconds =  9.89 MB/sec
/dev/hda5: Timing buffered disk reads:  64 MB in  5.08 seconds = 12.60 MB/sec
/dev/hda7: Timing buffered disk reads:  64 MB in  5.10 seconds = 12.55 MB/sec

2) 2.4 w/apm

/dev/hda1: Timing buffered disk reads:  64 MB in 16.03 seconds =  3.99 MB/sec
/dev/hda3: Timing buffered disk reads:  64 MB in 15.87 seconds =  4.03 MB/sec
/dev/hda4: Timing buffered disk reads:  64 MB in 15.67 seconds =  4.08 MB/sec
/dev/hda5: Timing buffered disk reads:  64 MB in 15.82 seconds =  4.05 MB/sec
/dev/hda7: Timing buffered disk reads:  64 MB in 15.85 seconds =  4.04 MB/sec

3) 2.4 w/o apm

/dev/hda1: Timing buffered disk reads:  64 MB in 16.02 seconds =  4.00 MB/sec
/dev/hda3: Timing buffered disk reads:  64 MB in 15.88 seconds =  4.03 MB/sec
/dev/hda4: Timing buffered disk reads:  64 MB in 15.67 seconds =  4.08 MB/sec
/dev/hda5: Timing buffered disk reads:  64 MB in 15.85 seconds =  4.04 MB/sec
/dev/hda7: Timing buffered disk reads:  64 MB in 15.86 seconds =  4.04 MB/sec   

Mine is on a ide1 device on a Dell Insp. 7500:
hdparm -i /dev/hda (from 2.2.17)

/dev/hda:

Model=IBM-DARA-225000, FwRev=SHAOA54A, SerialNo=SQASQ202564 Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs } RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=4 BuffType=3(DualPortCache), BuffSize=418kB, MaxMultSect=16, MultSect=off DblWordIO=no, OldPIO=2, DMA=yes, OldDMA=2
CurCHS=17475/15/63, CurSects=16513875, LBA=yes, LBAsects=49577472 tDMA={min:120,rec:120}, DMA modes: mword0 mword1 mword2 IORDY=on/off, tPIO={min:240,w/IORDY:120}, PIO modes: mode3 mode4 UDMA modes: mode0 mode1 *mode2 mode3 mode4
Drive Supports : ATA/ATAPI-4 T13 1153D revision 17 : ATA-1 ATA-2 ATA-3 ATA-4

...

Linda A Walsh                    | Trust Technology, Core Linux, SGI
[email protected]                      | Voice: (650) 933-5338
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
Please read the FAQ at http://www.tux.org/lkml/

[snip]

[snip]

Your slowdown's different from mine, then. hdparm gives me 16.5 MB/sec (or so) for both 2.2 & 2.4 (with APM in both cases). I must stress that I haven't tried 2.4.0 or 2.4.1pre on my Inspiron 5000e yet, so I don't even know if the problem that caused my slowdown still exists.

(Just to clarify my case: 750MHz Pentium III, 256MB RAM, 32GB IBM HD, Intel 440BX chipset. If you need more data, I'll collect it when I get a chance, but I'd be surprised if 2.4.1 doesn't come out first.)

While I'm writing this email, I may as well mention that I have a problem with booting my Inspiron 5000e under 2.4 (all of my actual 2.4test usage on that machines has been with Slackware 7.1's Disc 2 and the corresponding rootdisk). That machine has Red Hat 7.0 installed on an ext2 disk image file, stored on a vfat partition and loop mounted. (Actually, /usr and a subdirectory of /home are each on different image files now, for 3 total, but booting never gets far enough for that to be relevant.)

The RAM disk I use for booting off the hard drive is the default one set up by the Red Hat 7 installer, except there are no modules on it (I compile in everything that's needed for booting) and the last line of linuxrc has an echo statement. After that echo statement runs, I get the following error message under 2.4.0, but not 2.2.1[678]: (from memory, so the exact message might be slightly different)

panic: I have no root and I must scream

If I comment out that panic, I get some rather interestng messages, but I don't remember any of them. I wasn't planning on reporting this until I got a chance to put together a self-contained package & instructions for reproducing the problem, but since I'm writing an email to the list, here it is...

-Barry K. Nathan <[email protected]>
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
Please read the FAQ at http://www.tux.org/lkml/

Same problem here. I had a huge HDD performance drop when upgrading
from 2.2.18 to 2.4.0
It's an Intel i815 motherboard, and the HDD is Ultra-ATA.

Florin Andrei
"Saying everything is a database is saying nothing at all
and certainly will not improve communication with others.
Database, database database. Database! See?" (Marc Lehmann)
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
Please read the FAQ at http://www.tux.org/lkml/

ER, were you getting UDMA-100-66 out of 2.2.18 stock? Now what are you getting in 2.4.0?

Andre Hedrick
Linux ATA Development

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
Please read the FAQ at http://www.tux.org/lkml/

Mine was actually out of a stock 2.2.17 -- I tried your patch in an attempt to fix a disk problem - but the disk was just going bad and the slow speeds were coming from the automatic sector remapping.

pardon my ignorance, but where do you get UDMA-100-66?

Here is the hdparm -i output on 2.4:

/dev/hda:

Model=IBM-DARA-225000, FwRev=SHAOA54A, SerialNo=SQASQ202564 Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs } RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=4 BuffType=3(DualPortCache), BuffSize=418kB, MaxMultSect=16, MultSect=off DblWordIO=no, OldPIO=2, DMA=yes, OldDMA=2
CurCHS=17475/15/63, CurSects=16513875, LBA=yes, LBAsects=49577472 tDMA={min:120,rec:120}, DMA modes: mword0 mword1 mword2 IORDY=on/off, tPIO={min:240,w/IORDY:120}, PIO modes: mode3 mode4 UDMA modes: mode0 mode1 *mode2 mode3 mode4
Drive Supports : ATA/ATAPI-4 T13 1153D revision 17 : ATA-1 ATA-2 ATA-3 ATA-4

UDMA mode (2) seems to be identical to before.

Linda A Walsh                    | Trust Technology, Core Linux, SGI
[email protected]                      | Voice: (650) 933-5338
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
Please read the FAQ at http://www.tux.org/lkml/

Well:

i815 is an Ultra DMA 100 SouthBridge core.

Also, if your drive is caught in the question state of where the standard changes, the validity bits for determining the host/drive detection pair for the presense willl be fuzzy as is my explaination.

Andre Hedrick
Linux ATA Development

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
Please read the FAQ at http://www.tux.org/lkml/

Just as a datapoint, my Via IDE chipset (on atb850/kt133) and Promise Ultra66 (on 2xpp200/82440FX/PIIX3) works fine with 2.4.0, getting speeds about correct:

Model=IBM-DTLA-307045, FwRev=TX6OA60A
[snip. unused info cut]
PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 *udma4 udma5

 Timing buffered disk reads:  64 MB in  1.78 seconds = 35.96 MB/sec
 Timing buffered disk reads:  64 MB in  1.79 seconds = 35.75 MB/sec
 Timing buffered disk reads:  64 MB in  1.79 seconds = 35.75 MB/sec

Model=QUANTUM FIREBALLP LM30, FwRev=A35.0700
[snip. unused info cut]
tDMA={min:120,rec:120}, DMA modes: mword0 mword1 mword2 IORDY=on/off, tPIO={min:120,w/IORDY:120}, PIO modes: mode3 mode4 UDMA modes: mode0 mode1 mode2 mode3 *mode4

 Timing buffered disk reads:  64 MB in  4.30 seconds = 14.88 MB/sec
 Timing buffered disk reads:  64 MB in  4.34 seconds = 14.75 MB/sec

Quite slow for the drive is capable of, but it's very fast for the computer. All I care for that machine is it to be able to sustain a 100Mbit file transfer to it at near full speed :)

Some things I did notice is that sometimes I forget to mark down some of the options (dma especially) while compiling the kernel - and this killed my disk transfer rate...

(Also something weird: my Fujitsu just doesn't seem to want to get into UDMA66 (stays in udma33) in linux, but its head read rate is maxed out at 19MB/sec anyway, so it doesn't matter. It's also on the same IDE channel as the IBM on the Via. I really wonder why it's:

Model=FUJITSU MPD3137AH, FwRev=DH-05-09
[snip. unused info cut]
PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 mdma2 udma0 udma1 *udma2 udma3 udma4 Timing buffered disk reads: 64 MB in 3.23 seconds = 19.81 MB/sec

Maybe a firmware bug...
)

ydsmv

-bc

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
Please read the FAQ at http://www.tux.org/lkml/

<snip>

Hi,

I have the same drive, but "hdparm -i" gives different output:

Model=QUANTUM FIREBALLP LM30, FwRev=A35.0700, SerialNo=186006831114 Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs } RawCHS=16383/16/63, TrkSize=32256, SectSize=21298, ECCbytes=4 BuffType=DualPortCache, BuffSize=1900kB, MaxMultSect=16, MultSect=off CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=58633344 IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120} PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 *udma4

There is no mention of "UDMA modes."

However, my throughput seems to be much faster than yours for some reason:

Timing buffer-cache reads: 128 MB in 0.91 seconds =140.66 MB/sec
Timing buffered disk reads: 64 MB in 2.49 seconds = 25.70 MB/sec

My machine is an Athlon (booting says "Detected 848.381 MHz processor"). I'm not sure what the system bus speed is, but I think it's ~100MHz.

00:01.0 PCI bridge: Advanced Micro Devices [AMD] AMD-751 [Irongate] AGP Bridge (rev 01) (prog-if 00 [Normal decode])
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr-
Stepping- SERR+ FastB2B-
Status: Cap- 66Mhz+ UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort-
<MAbort- >SERR- <PERR-

	Latency: 64
	Bus: primary=00, secondary=01, subordinate=01, sec-latency=64
	I/O behind bridge: 0000e000-0000efff
	Memory behind bridge: fca00000-feafffff
	Prefetchable memory behind bridge: e4800000-f48fffff
	BridgeCtl: Parity- SERR+ NoISA- VGA+ MAbort- >Reset- FastB2B-

00:07.1 IDE interface: Advanced Micro Devices [AMD] AMD-756 [Viper] IDE (rev 03) (prog-if 8a [Master SecP PriP])
Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr-
Stepping- SERR- FastB2B-
Status: Cap- 66Mhz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort-
<MAbort- >SERR- <PERR-

	Latency: 64
	Region 4: I/O ports at cb00 [size=16]

CONFIG_BLK_DEV_IDEPCI=y
CONFIG_IDEPCI_SHARE_IRQ=y
CONFIG_BLK_DEV_IDEDMA_PCI=y
CONFIG_IDEDMA_PCI_AUTO=y
CONFIG_BLK_DEV_IDEDMA=y
CONFIG_BLK_DEV_AMD7409=y
CONFIG_IDEDMA_AUTO=y
CONFIG_BLK_DEV_IDE_MODES=y

Cheers,

Miles

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
Please read the FAQ at http://www.tux.org/lkml/

Hi!

Vmware is _not_ application, it is dirty kernel hack. Do it without vmware.
Pavel

I'm [email protected]. "In my country we have almost anarchy and I don't care."
Panos Katsaloulis describing me w.r.t. patents at [email protected]
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
Please read the FAQ at http://www.tux.org/lkml/