Re: [BENCHMARK] 2.4.{18,19{-ck9},20rc1{-aa1}} with contest

Dieter Nützel (Dieter.Nuetzel@hamburg.de)
Sat, 9 Nov 2002 04:44:44 +0100

Messages sorted by: [ date ][ thread ][ subject ][ author ]
Next message: Dieter Nützel: "Re: Linux v2.5.46 (compile error) (rivafb)"
Previous message: Matthew Wilcox: "Re: [parisc-linux] Untested port of parisc_device to generic device interface"
Maybe in reply to: Con Kolivas: "[BENCHMARK] 2.4.{18,19{-ck9},20rc1{-aa1}} with contest"
Next in thread: Con Kolivas: "Re: [BENCHMARK] 2.4.{18,19{-ck9},20rc1{-aa1}} with contest"
Reply: Con Kolivas: "Re: [BENCHMARK] 2.4.{18,19{-ck9},20rc1{-aa1}} with contest"

Andreww Morton wrote:
> Con Kolivas wrote:
> >
> > io_load:
> > Kernel [runs] Time CPU% Loads LCPU% Ratio
> > 2.4.18 [3] 474.1 15 36 10 6.64
> > 2.4.19 [3] 492.6 14 38 10 6.90
> > 2.4.19-ck9 [2] 140.6 49 5 5 1.97
> > 2.4.20-rc1 [2] 1142.2 6 90 10 16.00
> > 2.4.20-rc1aa1 [1] 1132.5 6 90 10 15.86
> >
>
> 2.4.20-pre3 included some elevator changes. I assume they are the
> cause of this. Those changes have propagated into Alan's and Andrea's
> kernels. Hence they have significantly impacted the responsiveness
> of all mainstream 2.4 kernels under heavy writes.
>
> (The -ck patch includes rmap14b which includes the read-latency2 thing)

No, the 2.4.19-ck9 that I have (the default?) include -AA and preemption (!!!)

Preemption is for several months the clear throughput winner for me.
Latest 2.4.19-ck9 and now 2.5.46-mm1.

I know you all "hate" dbench but 2.5.45/2.5.46-mm1 halved (!!!) my "dbench 32"
numbers. deadline IO is so GREAT.

2.4.19-ck5: ~55-60 seconds
2.5.46-mm1: ~31-45 seconds (even under VM pressure)

total used free shared buffers cached
Mem: 1034988 864172 170816 0 231840 345120
-/+ buffers/cache: 287212 747776
Swap: 1028120 8452 1019668
Total: 2063108 872624 1190484

Throughput 110.61 MB/sec (NB=138.263 MB/sec 1106.1 MBit/sec)
7.941u 38.251s 0:39.20 117.8% 0+0k 0+0io 841pf+0w

Sorry, "free -t" forgotten.

Throughput 114.462 MB/sec (NB=143.077 MB/sec 1144.62 MBit/sec)
7.986u 35.900s 0:37.90 115.7% 0+0k 0+0io 841pf+0w

total used free shared buffers cached
Mem: 1034988 481812 553176 0 178788 54048
-/+ buffers/cache: 248976 786012
Swap: 1028120 9836 1018284
Total: 2063108 491648 1571460

Throughput 112.283 MB/sec (NB=140.354 MB/sec 1122.83 MBit/sec)
7.728u 37.358s 0:38.62 116.7% 0+0k 0+0io 841pf+0w

total used free shared buffers cached
Mem: 1034988 461736 573252 0 163260 51488
-/+ buffers/cache: 246988 788000
Swap: 1028120 9976 1018144
Total: 2063108 471712 1591396

Only one MP3 playback hiccup during "dbench 32" and nearly no slowdown of
dbench.

2.5.45+ need some more memory during my normal workload and do little more
swap than 2.4.19+AA.

MemTotal: 1034988 kB
MemFree: 559784 kB
MemShared: 0 kB
Buffers: 164260 kB
Cached: 63308 kB
SwapCached: 2884 kB
Active: 399388 kB
Inactive: 10096 kB
HighTotal: 131008 kB
HighFree: 46508 kB
LowTotal: 903980 kB
LowFree: 513276 kB
SwapTotal: 1028120 kB
SwapFree: 1018156 kB
Dirty: 44 kB
Writeback: 0 kB
Mapped: 220700 kB
Slab: 36904 kB
Committed_AS: 530908 kB
PageTables: 3436 kB
ReverseMaps: 125959
HugePages_Total: 0
HugePages_Free: 0
Hugepagesize: 4096 kB

slabinfo - version: 1.2
fib6_nodes 7 112 32 1 1 1 : 248 124
ip6_dst_cache 9 20 192 1 1 1 : 248 124
ndisc_cache 1 30 128 1 1 1 : 248 124
raw6_sock 0 0 576 0 0 1 : 120 60
udp6_sock 1 7 576 1 1 1 : 120 60
tcp6_sock 5 8 1024 2 2 1 : 120 60
ip_conntrack 8 60 320 5 5 1 : 120 60
unix_sock 192 261 448 29 29 1 : 120 60
tcp_tw_bucket 0 0 128 0 0 1 : 248 124
tcp_bind_bucket 13 112 32 1 1 1 : 248 124
tcp_open_request 0 0 128 0 0 1 : 248 124
inet_peer_cache 2 59 64 1 1 1 : 248 124
secpath_cache 0 0 32 0 0 1 : 248 124
flow_cache 0 0 64 0 0 1 : 248 124
xfrm4_dst_cache 0 0 192 0 0 1 : 248 124
ip_fib_hash 15 112 32 1 1 1 : 248 124
ip_dst_cache 25 100 192 5 5 1 : 248 124
arp_cache 3 60 128 2 2 1 : 248 124
raw4_sock 0 0 448 0 0 1 : 120 60
udp_sock 7 18 448 2 2 1 : 120 60
tcp_sock 24 40 896 10 10 1 : 120 60
sgpool-MAX_PHYS_SEGMENTS 32 33 2560 11 11 2 : 54 27
sgpool-64 32 33 1280 11 11 1 : 54 27
sgpool-32 32 36 640 6 6 1 : 120 60
sgpool-16 32 36 320 3 3 1 : 120 60
sgpool-8 36 40 192 2 2 1 : 248 124
reiser_inode_cache 3900 19320 384 1932 1932 1 : 120 60
eventpoll 0 0 96 0 0 1 : 248 124
kioctx 0 0 192 0 0 1 : 248 124
kiocb 0 0 192 0 0 1 : 248 124
dnotify_cache 0 0 20 0 0 1 : 248 124
file_lock_cache 104 160 96 4 4 1 : 248 124
fasync_cache 2 202 16 1 1 1 : 248 124
shmem_inode_cache 12 27 448 3 3 1 : 120 60
uid_cache 5 112 32 1 1 1 : 248 124
deadline_drq 1792 1792 32 16 16 1 : 248 124
blkdev_requests 1280 1320 192 66 66 1 : 248 124
biovec-BIO_MAX_PAGES 256 260 3072 52 52 4 : 54 27
biovec-128 256 260 1536 52 52 2 : 54 27
biovec-64 256 260 768 52 52 1 : 120 60
biovec-16 256 260 192 13 13 1 : 248 124
biovec-4 256 295 64 5 5 1 : 248 124
biovec-1 325 404 16 2 2 1 : 248 124
bio 272 295 64 5 5 1 : 248 124
sock_inode_cache 237 330 384 33 33 1 : 120 60
skbuff_head_cache 897 980 192 49 49 1 : 248 124
sock 7 10 384 1 1 1 : 120 60
proc_inode_cache 117 696 320 58 58 1 : 120 60
sigqueue 87 87 132 3 3 1 : 248 124
radix_tree_node 4560 11340 320 945 945 1 : 120 60
cdev_cache 24 177 64 3 3 1 : 248 124
bdev_cache 15 30 128 1 1 1 : 248 124
mnt_cache 24 59 64 1 1 1 : 248 124
inode_cache 548 588 320 49 49 1 : 120 60
dentry_cache 7302 36560 192 1828 1828 1 : 248 124
filp 2512 2550 128 85 85 1 : 248 124
names_cache 6 6 4096 6 6 1 : 54 27
buffer_head 56609 158616 52 2203 2203 1 : 248 124
mm_struct 90 110 384 11 11 1 : 120 60
vm_area_struct 5357 6300 128 210 210 1 : 248 124
fs_cache 90 295 64 5 5 1 : 248 124
files_cache 90 99 448 11 11 1 : 120 60
signal_act 99 99 1344 33 33 1 : 54 27
task_struct 133 145 1600 29 29 2 : 54 27
pte_chain 19930 28851 64 489 489 1 : 248 124
mm_chain 0 0 8 0 0 1 : 248 124
size-131072(DMA) 0 0 131072 0 0 32 : 8 4
size-131072 0 0 131072 0 0 32 : 8 4
size-65536(DMA) 0 0 65536 0 0 16 : 8 4
size-65536 0 0 65536 0 0 16 : 8 4
size-32768(DMA) 0 0 32768 0 0 8 : 8 4
size-32768 1 1 32768 1 1 8 : 8 4
size-16384(DMA) 0 0 16384 0 0 4 : 8 4
size-16384 11 15 16384 11 15 4 : 8 4
size-8192(DMA) 0 0 8192 0 0 2 : 8 4
size-8192 5 9 8192 5 9 2 : 8 4
size-4096(DMA) 0 0 4096 0 0 1 : 54 27
size-4096 198 212 4096 198 212 1 : 54 27
size-2048(DMA) 0 0 2048 0 0 1 : 54 27
size-2048 190 206 2048 99 103 1 : 54 27
size-1024(DMA) 0 0 1024 0 0 1 : 120 60
size-1024 268 268 1024 67 67 1 : 120 60
size-512(DMA) 0 0 512 0 0 1 : 120 60
size-512 512 512 512 64 64 1 : 120 60
size-256(DMA) 0 0 256 0 0 1 : 248 124
size-256 360 360 256 24 24 1 : 248 124
size-192(DMA) 0 0 192 0 0 1 : 248 124
size-192 54 60 192 3 3 1 : 248 124
size-128(DMA) 0 0 128 0 0 1 : 248 124
size-128 923 1050 128 35 35 1 : 248 124
size-64(DMA) 0 0 64 0 0 1 : 248 124
size-64 1851 2124 64 36 36 1 : 248 124
size-32(DMA) 0 0 64 0 0 1 : 248 124
size-32 1891 2065 64 35 35 1 : 248 124
kmem_cache 112 128 120 4 4 1 : 248 124

GREAT work!

Regards,
Dieter

-- Dieter Nützel Graduate Student, Computer Science

University of Hamburg Department of Computer Science @home: Dieter.Nuetzel at hamburg.de (replace at with @) - 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/

Next message: Dieter Nützel: "Re: Linux v2.5.46 (compile error) (rivafb)"
Previous message: Matthew Wilcox: "Re: [parisc-linux] Untested port of parisc_device to generic device interface"
Maybe in reply to: Con Kolivas: "[BENCHMARK] 2.4.{18,19{-ck9},20rc1{-aa1}} with contest"
Next in thread: Con Kolivas: "Re: [BENCHMARK] 2.4.{18,19{-ck9},20rc1{-aa1}} with contest"
Reply: Con Kolivas: "Re: [BENCHMARK] 2.4.{18,19{-ck9},20rc1{-aa1}} with contest"