* 2.4.10pre12aa1 pgbench
@ 2001-09-20 11:29 Magnus Naeslund(f)
2001-09-20 18:27 ` Andrea Arcangeli
0 siblings, 1 reply; 2+ messages in thread
From: Magnus Naeslund(f) @ 2001-09-20 11:29 UTC (permalink / raw)
To: Andrea Arcangeli; +Cc: linux-kernel
This is not any scientific super bench, but this is what matters for me :)
It's an Alpha UX-164 633mhz with 1.25GB memory.
It's considerably slower than 2.4.5aa1, sorry to say.
This is my pgbench diffs:
--- 2.4.5aa1.txt Thu Sep 20 09:13:30 2001
+++ 2.4.10pre12aa1.txt Thu Sep 20 13:08:37 2001
@@ -9,9 +9,9 @@
number of clients: 1
number of transactions per client: 10
number of transactions actually processed: 10/10
-tps = 25.139082(including connections establishing)
-tps = 28.741744(excluding connections establishing)
-0.02user 0.00system 0:00.80elapsed 3%CPU (0avgtext+0avgdata 0maxresident)k
+tps = 9.923214(including connections establishing)
+tps = 10.497442(excluding connections establishing)
+0.02user 0.00system 0:01.15elapsed 2%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+0minor)pagefaults 0swaps
@@ -24,9 +24,9 @@
number of clients: 1
number of transactions per client: 100
number of transactions actually processed: 100/100
-tps = 27.289761(including connections establishing)
-tps = 27.678849(excluding connections establishing)
-0.06user 0.05system 0:03.86elapsed 3%CPU (0avgtext+0avgdata 0maxresident)k
+tps = 13.779075(including connections establishing)
+tps = 13.875670(excluding connections establishing)
+0.05user 0.04system 0:07.42elapsed 1%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+0minor)pagefaults 0swaps
@@ -39,9 +39,9 @@
number of clients: 20
number of transactions per client: 10
number of transactions actually processed: 200/200
-tps = 17.166296(including connections establishing)
-tps = 17.917905(excluding connections establishing)
-0.11user 0.14system 0:11.87elapsed 2%CPU (0avgtext+0avgdata 0maxresident)k
+tps = 16.683457(including connections establishing)
+tps = 17.374824(excluding connections establishing)
+0.11user 0.13system 0:12.17elapsed 2%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+0minor)pagefaults 0swaps
@@ -54,9 +54,9 @@
number of clients: 20
number of transactions per client: 100
number of transactions actually processed: 2000/2000
-tps = 27.149347(including connections establishing)
-tps = 27.330001(excluding connections establishing)
-0.98user 1.27system 1:13.97elapsed 3%CPU (0avgtext+0avgdata 0maxresident)k
+tps = 26.025792(including connections establishing)
+tps = 26.189368(excluding connections establishing)
+0.84user 1.33system 1:17.03elapsed 2%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+0minor)pagefaults 0swaps
@@ -69,9 +69,9 @@
number of clients: 50
number of transactions per client: 10
number of transactions actually processed: 500/500
-tps = 22.105250(including connections establishing)
-tps = 23.342988(excluding connections establishing)
-0.28user 0.40system 0:23.00elapsed 3%CPU (0avgtext+0avgdata 0maxresident)k
+tps = 19.077870(including connections establishing)
+tps = 19.985427(excluding connections establishing)
+0.28user 0.42system 0:26.57elapsed 2%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+0minor)pagefaults 0swaps
@@ -84,9 +84,9 @@
number of clients: 50
number of transactions per client: 100
number of transactions actually processed: 5000/5000
-tps = 20.743853(including connections establishing)
-tps = 20.863022(excluding connections establishing)
-2.55user 4.32system 4:01.46elapsed 2%CPU (0avgtext+0avgdata 0maxresident)k
+tps = 18.688064(including connections establishing)
+tps = 18.770892(excluding connections establishing)
+2.76user 4.20system 4:27.74elapsed 2%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+0minor)pagefaults 0swaps
@@ -99,9 +99,9 @@
number of clients: 100
number of transactions per client: 10
number of transactions actually processed: 1000/1000
-tps = 30.153119(including connections establishing)
-tps = 32.500791(excluding connections establishing)
-0.69user 1.11system 0:33.90elapsed 5%CPU (0avgtext+0avgdata 0maxresident)k
+tps = 16.874544(including connections establishing)
+tps = 17.579728(excluding connections establishing)
+0.64user 1.18system 0:59.94elapsed 3%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+0minor)pagefaults 0swaps
@@ -114,9 +114,9 @@
number of clients: 100
number of transactions per client: 100
number of transactions actually processed: 10000/10000
-tps = 16.258988(including connections establishing)
-tps = 16.322457(excluding connections establishing)
-6.59user 14.50system 10:15.27elapsed 3%CPU (0avgtext+0avgdata
0maxresident)k
+tps = 15.257859(including connections establishing)
+tps = 15.313045(excluding connections establishing)
+6.55user 14.54system 10:55.68elapsed 3%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (0major+0minor)pagefaults 0swaps
@@ -129,9 +129,9 @@
number of clients: 200
number of transactions per client: 10
number of transactions actually processed: 2000/2000
-tps = 20.326710(including connections establishing)
-tps = 21.324192(excluding connections establishing)
-1.62user 3.60system 1:39.77elapsed 5%CPU (0avgtext+0avgdata 0maxresident)k
+tps = 16.447976(including connections establishing)
+tps = 17.126929(excluding connections establishing)
+1.70user 3.72system 2:03.03elapsed 4%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+0minor)pagefaults 0swaps
@@ -144,7 +144,7 @@
number of clients: 200
number of transactions per client: 100
number of transactions actually processed: 20000/20000
-tps = 8.367056(including connections establishing)
-tps = 8.383456(excluding connections establishing)
-19.83user 73.54system 39:50.71elapsed 3%CPU (0avgtext+0avgdata
0maxresident)k
+tps = 7.374791(including connections establishing)
+tps = 7.387783(excluding connections establishing)
+19.79user 71.96system 45:15.92elapsed 3%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (0major+0minor)pagefaults 0swaps
Magnus
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Programmer/Networker [|] Magnus Naeslund
PGP Key: http://www.genline.nu/mag_pgp.txt
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: 2.4.10pre12aa1 pgbench
2001-09-20 11:29 2.4.10pre12aa1 pgbench Magnus Naeslund(f)
@ 2001-09-20 18:27 ` Andrea Arcangeli
0 siblings, 0 replies; 2+ messages in thread
From: Andrea Arcangeli @ 2001-09-20 18:27 UTC (permalink / raw)
To: Magnus Naeslund(f); +Cc: linux-kernel
On Thu, Sep 20, 2001 at 01:29:35PM +0200, Magnus Naeslund(f) wrote:
> This is not any scientific super bench, but this is what matters for me :)
>
> It's an Alpha UX-164 633mhz with 1.25GB memory.
> It's considerably slower than 2.4.5aa1, sorry to say.
I don't exclude it's a vm problem (and I know where to put the hands in
such case) but could you compare also with 2.4.10pre10 or anyways
something more recent than a 2.4.5 based kernel. There are been changes
in the I/O subsystem as well meanwhile (ask Jens). I will also need to
check what pgbench is doing, where can I get it?
Andrea
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2001-09-20 18:27 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2001-09-20 11:29 2.4.10pre12aa1 pgbench Magnus Naeslund(f)
2001-09-20 18:27 ` Andrea Arcangeli
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox