[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [MiNT] Libcurl-ssl / RPM package



I'm thinking its something to do with timing. Results from my Falcon, and then results from a Solaris 10 Sparc (which I happened to built OpenSSL 1.0.0a on today for my work):

falcon:apps# ./openssl speed blowfish
Doing blowfish cbc for 3s on 16 size blocks: 178717 blowfish cbc's in -603.00s
Doing blowfish cbc for 3s on 64 size blocks: 48671 blowfish cbc's in -600.00s
Doing blowfish cbc for 3s on 256 size blocks: 12457 blowfish cbc's in -600.00s
Doing blowfish cbc for 3s on 1024 size blocks: 3131 blowfish cbc's in -600.00s
Doing blowfish cbc for 3s on 8192 size blocks: 384 blowfish cbc's in -601.00s
OpenSSL 1.0.0a 1 Jun 2010
built on: Tue Jun 15 16:17:51 EDT 2010
options:bn(64,32) rc4(ptr,int) des(idx,cisc,2,long) aes(partial) idea(int) blowfish(idx)
compiler: m68k-atari-mint-gcc -DB_ENDIAN -DTERMIOS -m68020-60 -O3
The 'numbers' are in 1000s of bytes per second processed.
type             16 bytes     64 bytes    256 bytes   1024 bytes   8192 bytes
blowfish cbc     -4742.08     -5191.57     -5314.99     -5343.57     -5234.16
falcon:apps# uname -a
FreeMiNT falcon 1.16.3a 4.4 falcon mc68060
falcon:apps#
---
mfas-1n.lcs:lib# /usr/local/ssl/bin/openssl speed blowfish
Doing blowfish cbc for 3s on 16 size blocks: 7582343 blowfish cbc's in 2.92s
Doing blowfish cbc for 3s on 64 size blocks: 2068573 blowfish cbc's in 2.94s
Doing blowfish cbc for 3s on 256 size blocks: 527557 blowfish cbc's in 2.93s
Doing blowfish cbc for 3s on 1024 size blocks: 132321 blowfish cbc's in 2.91s
Doing blowfish cbc for 3s on 8192 size blocks: 13766 blowfish cbc's in 2.43s
OpenSSL 1.0.0a 1 Jun 2010
built on: Thu Jun 17 12:34:57 EDT 2010
options:bn(64,32) rc4(ptr,char) des(idx,cisc,16,long) aes(partial) idea(int) blowfish(ptr) compiler: gcc -fPIC -DOPENSSL_PIC -DOPENSSL_THREADS -D_REENTRANT -DDSO_DLFCN -DHAVE_DLFCN_H -m32 -mcpu=ultrasparc -O3 -fomit-frame-pointer -Wall -DB_ENDIAN -DBN_DIV2W -DOPENSSL_BN_ASM_MONT -DSHA1_ASM -DSHA256_ASM -DSHA512_ASM -DAES_ASM
The 'numbers' are in 1000s of bytes per second processed.
type             16 bytes     64 bytes    256 bytes   1024 bytes   8192 bytes
blowfish cbc     41547.08k    45030.16k    46093.72k    46562.44k    46407.85k
mfas-1n.lcs:lib#
---

If I peek at an 0.9.8 series on the Sparc, I see:

available timing options: TIMES TIMEB HZ=100 [sysconf value]
timing function used: times

So, it looks like the timing result is being multiplied by a -200 somehow.

Keith

On Thu, 17 Jun 2010, Howard Chu wrote:

Keith Scroggins wrote:
I just used:

openssl speed

That is builtin.

Then there's something broken in this part of the code. Either it's not reading the clock correctly, and so getting invalid timings, or there's a math overflow in there somewhere.

You should get results that look like this:

The 'numbers' are in 1000s of bytes per second processed.
type             16 bytes     64 bytes    256 bytes   1024 bytes   8192 bytes
md2 1738.22k 3547.29k 4792.66k 5272.06k 5368.49k

All of the numbers should be positive.

Possibly the code is so slow that it can't even complete one result in 3 seconds, but I highly doubt that. Something else is going on.

Keith

On Thu, 17 Jun 2010, Howard Chu wrote:

Keith Scroggins wrote:

Something is very wrong with all of the results posted, there should not be
negative numbers anywhere. What did you use to generate these tables?

--
 -- Howard Chu
 CTO, Symas Corp.           http://www.symas.com
 Director, Highland Sun     http://highlandsun.com/hyc/
 Chief Architect, OpenLDAP  http://www.openldap.org/project/