Posted January 31, 201411 yr Let's be gentlemen's and share our knowledge for this rather popular benchmark. Since I've started the thread, I will start with my share of tweaks I've been using (partially) for the last 2 years. Operating System * Preferred OS, in order of performance Linux MacOSX Windows NT5 (XP / 2003) Windows NT6 (Vista / 7 / 8 / 8.1) * General OS Tweaks For Windows platforms there are no specific requirements, other than having as few processes and services running. Do not try achieving this by using DIAGNOSTIC MODE, it will generally score lower than in normal mode with startup items disabled etc. Command Line Parameters * Using High Priority for unrar.exe You can achieve this in various ways (up to your general knowledge). Since 99.99% of the users are using Windows for this benchmark, they could use: start /high /b unrar bench test.rar Make sure that launcher.exe was started with Administrator rights for this to work. * Setting unrar.exe intensity unrar bench test.rar -ri# # is a number from 0 to 15 You can also set a sleep time (in milliseconds), which seems to also help unrar bench test.rar -ri# $ is a number representing the amount of milliseconds to sleep. It's up to you to test values, personally I've tried up to 10000. * Changing the Instruction Set It is important NOT to run using the "All Supported" option. General rule is to select the instruction set available before the newest supported. If your cpu supports SSE4, then SSE3 would be the preferred choice. The only exception here would be on CPUs with only SSE2 and i386, where you would definitely want to choose SSE2. Note that it's good practice to also test the latest instruction set, but it will rarely score any better. On AMD processors use XOP if available (Bulldozer and newer). Note that XOP only works under: Windows 7 SP1 / R2 SP2 / 8 / 8.1 OR Linux with kernels newer than 2.6.30 * Changing the Threads used THIS IS NO LONGER ALLOWED, NOTHING TO SEE HERE. ====================================================== UPDATE1: 07/SEP/2014 Thread changing removed. Edited September 7, 201410 yr by GENiEBEN
February 1, 201411 yr Same findings here. For the lazy guys: start /high /b unrar bench -cf=ssse3 -cpus=1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64 test.rar Don't forget to grab some popcorn.
February 1, 201411 yr Author And there's more... Of course there is, but there aren't as many gentleman's as expected.
February 1, 201411 yr I have this bench figured out Here is what can be done at 6ghz, damn I wish I had a chip that did easy 6.3ghz... There is helping someone tweak an OS then there is giving away the golden goose. I prefer the former!
February 1, 201411 yr Author Tweaking Windows to improve scores for UCBench is like removing the backseats from a Civic for acceleration improvement.
February 1, 201411 yr Crew There was a tweak for Windows7, alike Bob's tweak for PCMark05. In aero mode it gave me a score of a cpu running 50-100Mhz more... no idea bout you , but I always get better scores with SSE2 then SSE3... I demonstrated it to Massman in a former countrycup, have to see if this still works...
February 10, 201411 yr For the guys who have this bench figured out, what sort of score should I be looking to get for a 4.5GHz i7 4770K and what OS should I be using?
February 10, 201411 yr Author For the guys who have this bench figured out, what sort of score should I be looking to get for a 4.5GHz i7 4770K and what OS should I be using? See OP, if going for Windows, then a very lite XP would work best.
February 11, 201411 yr OS choise depents on socket thought. On X58 XP 64 seems the way to go. Lite Win is an bad choise, disable most services let score go down Will test Vista and Vista x64 tomorrow
February 12, 201411 yr Thanks for sharing Moose! I don't like this bench really, but it has opened up benching with some cool 1 and 2 core cpus.. and that's great
February 12, 201411 yr Yeah cheers for the advice moose. Was all set up to do some testing but very strong winds here has left us without power all day and it has only just come back on now. Hopefully I can get some prep work done tonight and then some testing tomorrow
February 13, 201411 yr ^ never tried that but seems you dont need any mem speed or cache speed on z87 Holy shit, I tried on win7 with the 1/2/3/4/5/6/7/etc all threads tweak, same MHz as you and only got 1181. Why the hell is the difference so immense?
February 14, 201411 yr Yeah I know he's on XP, but does it really make that much of a difference? I'll definitely try then
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.