* [Qemu-devel] [PATCH 0/6] A cumulative MIPS patchset
@ 2006-11-21 14:33 Thiemo Seufer
0 siblings, 0 replies; only message in thread
From: Thiemo Seufer @ 2006-11-21 14:33 UTC (permalink / raw)
To: qemu-devel
Hello All,
The following patchset collects the most important improvements of
MIPS system emulation mode:
qemu.drow-performance # TLB performance improvements
qemu.drow-halt-reboot # support for halt and reboot
moredevices # A few more devices, most importantly RTC
mfc0-inline # inline MFC0 instructions
mtc0-inline # inline MTC0 instructions
opcode-handling # MIPS32R2 support and retionalised instruction
# decoding
Daniel's improvements of the TLB emulation allow finally some useful
performance measurements for inlining of CP0 accesses. The appended
lmbench result shows general improvements. The first line is qemu
patched with the whole patchset, the second is qemu patched up to
moredevices. Both runs usd the same binaries, without using the
MIPS32R2 improvements.
Thiemo
L M B E N C H 3 . 0 S U M M A R Y
------------------------------------
(Alpha software, do not distribute)
Basic system parameters
------------------------------------------------------------------------------
Host OS Description Mhz tlb cache mem scal
pages line par load
bytes
--------- ------------- ----------------------- ---- ----- ----- ------ ----
lagash-qe Linux 2.6.18- mips-linux-gnu 185 1
lagash-qe Linux 2.6.18- mips-linux-gnu 185 1
Processor, Processes - times in microseconds - smaller is better
------------------------------------------------------------------------------
Host OS Mhz null null open slct sig sig fork exec sh
call I/O stat clos TCP inst hndl proc proc proc
--------- ------------- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
lagash-qe Linux 2.6.18- 185 1.48 4.52 53.0 85.0 126. 12.1 71.6 7104 33.K 142K
lagash-qe Linux 2.6.18- 185 2.05 5.71 55.0 91.6 135. 13.0 74.0 7221 33.K 142K
Basic integer operations - times in nanoseconds - smaller is better
-------------------------------------------------------------------
Host OS intgr intgr intgr intgr intgr
bit add mul div mod
--------- ------------- ------ ------ ------ ------ ------
lagash-qe Linux 2.6.18- 5.9500 0.0900 20.7 32.6 37.9
lagash-qe Linux 2.6.18- 6.0300 0.0900 20.9 33.4 38.5
Basic uint64 operations - times in nanoseconds - smaller is better
------------------------------------------------------------------
Host OS int64 int64 int64 int64 int64
bit add mul div mod
--------- ------------- ------ ------ ------ ------ ------
lagash-qe Linux 2.6.18-
lagash-qe Linux 2.6.18-
Basic float operations - times in nanoseconds - smaller is better
-----------------------------------------------------------------
Host OS float float float float
add mul div bogo
--------- ------------- ------ ------ ------ ------
lagash-qe Linux 2.6.18- 77.7 114.3 153.9 536.6
lagash-qe Linux 2.6.18- 78.6 113.9 155.7 543.0
Basic double operations - times in nanoseconds - smaller is better
------------------------------------------------------------------
Host OS double double double double
add mul div bogo
--------- ------------- ------ ------ ------ ------
lagash-qe Linux 2.6.18- 154.4 178.4 362.2 1071.8
lagash-qe Linux 2.6.18- 156.2 188.0 366.4 1078.5
Context switching - times in microseconds - smaller is better
-------------------------------------------------------------------------
Host OS 2p/0K 2p/16K 2p/64K 8p/16K 8p/64K 16p/16K 16p/64K
ctxsw ctxsw ctxsw ctxsw ctxsw ctxsw ctxsw
--------- ------------- ------ ------ ------ ------ ------ ------- -------
lagash-qe Linux 2.6.18- 49.1 127.1 380.8 291.7 692.9 343.4 796.5
lagash-qe Linux 2.6.18- 49.0 116.3 350.2 291.1 668.7 342.9 784.2
*Local* Communication latencies in microseconds - smaller is better
---------------------------------------------------------------------
Host OS 2p/0K Pipe AF UDP RPC/ TCP RPC/ TCP
ctxsw UNIX UDP TCP conn
--------- ------------- ----- ----- ---- ----- ----- ----- ----- ----
lagash-qe Linux 2.6.18- 49.1 199.6 409. 782.2 1940. 7957
lagash-qe Linux 2.6.18- 49.0 193.2 433. 851.3 1888. 7827
*Remote* Communication latencies in microseconds - smaller is better
---------------------------------------------------------------------
Host OS UDP RPC/ TCP RPC/ TCP
UDP TCP conn
--------- ------------- ----- ----- ----- ----- ----
lagash-qe Linux 2.6.18-
lagash-qe Linux 2.6.18-
File & VM system latencies in microseconds - smaller is better
-------------------------------------------------------------------------------
Host OS 0K File 10K File Mmap Prot Page 100fd
Create Delete Create Delete Latency Fault Fault selct
--------- ------------- ------ ------ ------ ------ ------- ----- ------- -----
lagash-qe Linux 2.6.18- 9447.0 11.8 41.2 80.8
lagash-qe Linux 2.6.18- 10.3K 13.7 46.4 123.1
*Local* Communication bandwidths in MB/s - bigger is better
-----------------------------------------------------------------------------
Host OS Pipe AF TCP File Mmap Bcopy Bcopy Mem Mem
UNIX reread reread (libc) (hand) read write
--------- ------------- ---- ---- ---- ------ ------ ------ ------ ---- -----
lagash-qe Linux 2.6.18- 28.7 26.7 14.2 49.1 85.3 66.7 44.0 87.5 110.6
lagash-qe Linux 2.6.18- 28.6 27.1 14.6 47.8 83.7 63.9 43.1 84.1 109.5
Memory latencies in nanoseconds - smaller is better
(WARNING - may not be correct, check graphs)
------------------------------------------------------------------------------
Host OS Mhz L1 $ L2 $ Main mem Rand mem Guesses
--------- ------------- --- ---- ---- -------- -------- -------
lagash-qe Linux 2.6.18- 185 16.9 21.0 574.1 13.5K No L1 cache?
lagash-qe Linux 2.6.18- 185 17.8 22.0 563.5 13.8K No L1 cache?
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2006-11-21 14:33 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2006-11-21 14:33 [Qemu-devel] [PATCH 0/6] A cumulative MIPS patchset Thiemo Seufer
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).