Existing QL benchmarks versus Q68

Nagging hardware related question? Post here!
User avatar
Peter
Gold Card
Posts: 427
Joined: Sat Jan 22, 2011 8:47 am

Existing QL benchmarks versus Q68

Postby Peter » Sun Nov 19, 2017 9:21 am

Benchmarks to compare hardware speed... are they right under hardware? Or software? General chat? Please correct me if this is the wrong location.

Some Dhrystone/s results from http://sinclairql.emuunlim.com/bench.htm:

Code: Select all

QXL  20MHz:                        6578
QXL2 25MHz:                        8620
Q40:                              36443
Q60/80:                          101010

For the Q68 I measured:

Code: Select all

Q68 (Mode 4 512x256)               6925

Some QSBB_BAS results from http://www.jadiam.org/QL/QLAY2:

Code: Select all

QL original       980       840    1100
GoldCard         2580      4280    5680
SuperGoldCard    5680     10260   13820
QXL 20MHz       14300     27900   70640

For the Q68 I measured:

Code: Select all

                print  function  string
Q68             20760     28700   63189

Some Linpack results from http://forum.tlienhard.com/phpBB3/viewtopic.php?f=9&t=836

Code: Select all

GoldCard            5.5 kFlop/s
SuperGoldCard      14.7 kFlop/s
QXL 20MHz          22   kFlop/s
Q60               228   kFlop/s

For the Q68 I measured:

Code: Select all

Q68                22.9 KFlop/s


This can not be an exact comparison, I didn't reproduce the results for other machines myself. Except for the Original QL, I assume that SMSQ/E and Mode 4 512x256 were used, cache activated where possible. More results welcome.

Peter


User avatar
Peter
Gold Card
Posts: 427
Joined: Sat Jan 22, 2011 8:47 am

Re: Existing QL benchmarks versus Q68

Postby Peter » Fri Nov 24, 2017 2:28 pm

Some QTop-Index v1.07 results from http://qlforum.co.uk/viewtopic.php?f=3&t=1500:

Code: Select all

SuperGoldCard with Aurora      13.770
Q68                            23.718
Q60                           195.892


User avatar
Zarchos
Brittle Membrane
Posts: 106
Joined: Mon May 08, 2017 11:49 am

Re: Existing QL benchmarks versus Q68

Postby Zarchos » Fri Nov 24, 2017 3:21 pm

Can you force some of your tests to run in the 'special' fast RAM area ?

Even when accessing screen memory for example (which isn't in the fast RAM area), running the code in the fast RAM area should give some speed gains,
as the decoding of the 68000 instructions should be much faster.



So you've got something very important to do right now ?
Why not first waste some precious minutes of your life watching some stupid videos about games and demos on the Archimedes on my YT channel ?
---> https://www.youtube.com/user/Archimedes75009
User avatar
Peter
Gold Card
Posts: 427
Joined: Sat Jan 22, 2011 8:47 am

Re: Existing QL benchmarks versus Q68

Postby Peter » Fri Nov 24, 2017 7:22 pm

Not quickly. In full length they don't fit. The Q68 was not designed for speed - the benchmarks are meant so potential Q68 buyers have realistic expectations and can make an informed decision.

And yes, screen access routines would also run significantly faster in Fast RAM.


FrancoisLanciault
ROM Dongle
Posts: 22
Joined: Mon Aug 08, 2011 11:08 pm

Re: Existing QL benchmarks versus Q68

Postby FrancoisLanciault » Sat Nov 25, 2017 1:42 am

Faster than SuperGoldCard, good enough for me !!!

Looking forward to play with this marvelous board.

François



Return to “Hardware”

Who is online

Users browsing this forum: No registered users and 5 guests