Taloken

Well S4ch4z result was out before Nexus submitted his better DDR3 I understand why Nexus' DDR3 didn't appear (he already have a better DDR4 score), but i don't see what crashed S4ch4z' DDR3 About 2) : Well Top-3 scores is too easy i think : For Stage 3 for example, 3 novices with Z97+CFR/MFR and it's done. By adding the hardware-type, and one score per user per stage, it encourage teams to dig for members and hardware ^^

Exotic, exotic ... yeah people eat snails and you call them exotics tsss Kidding, thx for the quick feedback ^^

S4ch4z's doesn't have DDR4 rig. Nexus (current DDR4 score *sigh*) submitted a better DDR3, but it doesn't appear neither his DDR4 score isn't removed.

Ok thx ! I ask other members for their log file to help you find a general case issue

Hi. S4ch4z's MaxxMem score disappeared of our results, however it remains linked to the comp : http://hwbot.org/submission/2835244_s4ch4z_maxxmem_read_bandwidth_ddr3_sdram_25650_mbytesec.

Hi ! We (2 teammates and me) can't submit any datafile. By automatic upload (with/without one/both skips detection), error message that datafile is maybe corrupted, By manual uploading, invalid datafile. I join you my log with a Parse-like Exception. After taking a look, my datafile is encoded in ANSI. As Hwbot is asking for UTF-8,i tried to reencode with npp++ butwithout success. ++

If you see some spaks, take sausages quickly to enjoy the extra volts

Thanks websmile and pieter ^^

You mean that 10 is not impacted by the RTC bug ? Oo Thx for the precisions

Hi all. First, for XTU, can the settings be private are they have to be public ? please explicitly precise it in bench's rules ^^ Also, in all concerned benchs rules, can you add W10 in the list of disallowed W8 based OS ? It's not explicit and some people are tented to use this consciously for submitting invalid scores. Thx !

"Capable of working under pressure" Encouraging.

GJ Shep, push it !

"Gigabyte B85M-D3V" : KISS-Approved ! (Keep It Simple Stupid)

Bugged runs/scores/detetcions happens on ALL benchmarks/tools, no need to "prove" it ...

LN2 and Bulldozers are too mainstream.

Yep keep 1 or 2 cores for Max Freq, it's easier for cpu to scale ^^ Also for single-threaded benchmarks like SuperPi, PiFast, you can disable cores to stress less the CPU. You can (allowed) also on multi-threaded, but it don't bring anything, just making the score lower x)

Godavari APU are announced for end-May/early-June, will they be allowed for this round ?

I know about 7/64 bits/AMD and other stuffs, but it's on my working PC so i can't just reinstall an OS for that. Anyway, i'll check about old Catalyst, thx Mat

Hi ! I want to gpupi a Pentium G620 on a Windows XP 32 bits SP3. AMD APP SDK don't support XP, and with Intel Runtime 15.1, i have an "OpenCL.dll not found"-like error. I searched in the file system for the dll but didn't find it. What driver can work with XP ?

Yeah sometimes it takes time for submissions to be included in ranking. Go on your submit page, and add "?recalculating=true" at the end of the URL, then Enter. Do it while points are not updated. There is a "Force recalculation" in the right panel, but don't know if it really workd.

Seems like sharing is hard with local products on stage 3 ....

You have a "edit" button on the top of right panel

I also don't see the point of this. Ones who bench only for competitions will let Extreme-cooling aside, as watercooling is cheaper and easier to manage (full load on fans, no CB/CBB and go on). Also, the main goal of overclocking is bringing technology at his limits. An Ambient World Record would not have many senses, as it don't break a new wall. Without LN2, how would we catch the "super-ultra-novator TIM" of Devil's Canyon as crap ? I'm Enthusiast too, not thinking of LN2 for the moment, but i take pleasure to compete versus Extremes in rankings. Actually, with my FX, i try to reach top scores of Ambient-cooling, and even if it give me only 2~3 HW+GL Points, there is challenge. (Hey, what about a new cooling-filter in rankings Ambient/Extreme ?) Also competitions are more and more interested in "low-cost clocking", take example with ROG OC Formula, or the use of the G3240 in the G.Skill World Cup.

Yep but for NN#2, you make the GpuPi stage as a per-core stage. http://forum.hwbot.org/showthread.php?t=132135 In my memories, all multi-threaded benchmarks for RR (hwbot prime, cinebench ...) were implemented in a per-core stage.

Well, about tweaking of any bench, we heard white on one side, black on an other, rarely these bloody shades of grey ... Sometimes i have the feeling that some tweaks are just placebo effects I will give a try thx ^^