I.M.O.G.

I kinda feel like this is a solution looking for a problem. Isn't it the natural order of things in benchmarking that your scores do not stand forever? The stronger your score, the longer it is likely to stand. I don't see a problem in that... If you want to keep your rankings, you have to stay up to date to defend them. I like that. If there is a perceived problem with that, I think quarterly ranking snapshots would be cool so you can compare to last quarter, this quarter last year, and a quarter maybe 4 years ago just to look at how things changed and the trends. It would be neat to see a capture like that of what the dominant platforms looked like over a timeline... Currently that is lost as older stuff is trampled. I think snapshots would be a better answer than multiplying the number of rankings further. Sure, the snapshots wouldn't count for points in the current rankings, but people would still have the honor of being captured at the top of past snapshots and that's something cool to strive for (something worth sandbagging for even perhaps?). The more divisions there are in the rankings, the harder the rankings are to police, and there is already enough garbage in the rankings that is hard to detect and eliminate at non-top levels (that isn't a criticism of hwbot, just recognition of a difficult problem to address given the number of submissions).

I think massman commented before, and the high core count point rewards in pcmark05 were kind of an unexpected side effect of the rev4 point scheme. It was sort of highlighted when 1st place in the enthusiast league was taken rather easily from just running high core count benches at stock - I don't remember where exactly he posted about it, but it was a conversation related to that. Kinda cool to see what those sorts of rigs are capable of, but the points rewarded don't reflect the effort involved. There needs to be some sort of split rankings for pcmark however, otherwise no one would want to bother running single core, dual core, or other configs for top rankings... Those categories are pretty intensely competitive currently. The top 50 rankings for pcmark would be almost exclusively sandybridge chips if there was no split, making running anything else for the bench almost pointless.

Godly. Seriously though: Areca 1882

The 555 did fine, and lived to tell the story. Got a slightly better pifast out of it than the 8120: http://hwbot.org/submission/2260783 PCMark05 did 43K: http://hwbot.org/submission/2260786 Good scores, probably not good enough to lock up the competition though with 14 days left - guess we'll find out 5 minutes before the competition ends. I've had one better pifast and wprime run, and 3 better superpi32m scores - all on chips I no longer own however. PCMark05 is the only one I set a new personal best on, I'd feel safer if I broke my own records on the other ones.

Look at the global rankings... The Sandy Bridge has me surrounded!

The rerun of my 42K PCM05 score won't happen (I forgot to use the background on that one)... My golden FX-8120 just died at stock in the 990FXA-UD7. Going to give it a shot at killing my 555 BE now as well, fingers crossed. Should put up about the same score anyways. Losing that 8120 stings though, it was good while it lasted: http://hwbot.org/submission/2225833_i.m.o.g._cpu_frequency_fx_8120_8235.25_mhz

Solid storage scores as always Steve. Nice work.

Very funny.

It's essentially the same tweak I think Sweet, except without some steps that don't actually make any difference as far as I can tell. Sebas was working his butt off to try to figure it out while people gave hints, so I think thats how he found his own way of doing it. I'd trade you my method for that Revodrive though.

Pifast isn't very tweakable, it's almost purely CPU frequency based. Your ram speed and timing makes a minimal difference - just change your ram divider from high to low and keep the CPU freq the same to see what I mean, no big difference. sp32m is the opposite. cpu frequency is still the most important, but the memory subsystem and ram itself are very important to good scores as well. Don't forget to load graphics drivers too.

Whoa! Nice stuff.

No, the version of the transparent window tweak I apply doesn't involve logging off or modifying appearance settings. I just checked the motherboard - I used a new insulation approach this time, and it looks like its dried off from last night. I should be able to post a valid gigabyte competition submission tonight, and I'm hoping to hit 44K.

I wish you were entitled to take your whining to a thread dedicated to that purpose. Only on HWBot is there virtually no thread moderation and regular members are allowed to persistently troll a thread pushing their own agenda, without contributing anything on topic. If a thread is titled "PCMark sucks get rid of it", I'm going to read it once then probably not look again. If a thread is titled "PCMark05 tweaking thread" I'm going to keep track of its updates. Every site has its own opinion about what level of moderation is appropriate. HWBot is very laissez-faire, and this is a perfect example why sometimes, in my opinion, it is good to have moderators who are willing to put a boot in someone's ass so that those with a genuine interest in a thread are not trolled constantly. No offense Duke, but how about starting your own thread instead of using this tweaking and sharing thread as your personal dumping ground. You have an agenda to get rid of PCMark, we get it... Just use your own thread, or another appropriate thread. This thread isn't the place for it is all.

Yup, Bulldozer is just slow in these benchmarks. Heard good things about it for Battlefield3 which is well threaded, but in these benches it blows. If you look at my past submissions with the FX-8120, they are better scores... My best SP32M on Bulldozer was 10M 53S @ 7.74GHz. I was also using stronger RAM and CPU-NB for BD, but it only beat my best PhenomII SP32M by 2 seconds. There aren't many people that have a BD chip capable of running the clocks I'm pulling, and its just barely competitive... Thuban is the chip to have in this competition. I'll be making more submissions yet. The pifast score will be improved without much trouble, and I have to fix the PCMark run since I forgot to use the official background. The SP32M run could be better, but I probably won't mess with it... wPrime32M is as good as I can do. The PCMark05 and wprime32m scores are the highest 8 core AMD scores submitted to hwbot so far though.

Not bob's tweak exactly, but more or less the same thing. Did this for the gigabyte competition, but forgot to use the official background. Need to rerun it and submit for that following the rules correctly - just submitted it regular until I get it rerun.

The Jiggy mobo treated me alright tonight. Got 42K, but need to rerun because I forgot to use the background. Next goal is 43K I guess: http://hwbot.org/submission/2259791_i.m.o.g._pcmark_2005_fx_8120_42727_marks

Rerun it is then. 43K here we come... Is a mod able to remove this submission from the competition, without deleting it? Or should I just resubmit as a non-competition entry? EDIT: Deleted and resubmitted.

Ya, I mean I can rerun if I need to, but would prefer not to waste the extra LN2. The rules should be accessible from the main competition page. I didn't realize there as a desktop background requirement until I actually accessed one of the stages directly. Noob mistake on my part, I don't normally take part in official vendor competitions.

Forgot to use the competition background - should I rerun this?

I didn't see that comment about the 7970. It isn't a big advantage - the 7970 is the best card for pcm05, but it doesn't make a big difference. It does make an important difference if you are going for global first place where everyone elses scores are setting records, but in a competition like this where records may not be broken there are other factors that can be more important. For instance, I should beat his physics/3d and be neck and neck with his 2d graphics score, but his 3d pixel shader is better than I can possibly do... I'll be using a 5870. My target PCMark05 score is 42K, so that I can reclaim 8 core 1st place global. If I don't hit that, I should at least be in the 40K range, unless I have problems getting the right clocks on this board. FTW's pcmark score has everything he needs, except for stronger storage performance. His score is excellent without that storage, and I expect he'll do higher yet. I'll be doing some runs shortly.

I could be wrong on that, I thought I had seen single sata3 ssds do 220. I've only done revodrive and raid0. I have a crappy sata3 drive sitting here that does just about 200 every time tho. I will be joining. I sell my hardware to buy new hardware though, so I will have a hard time competing in sp32m and pifast on bulldozer. To be in the ballpark, I will have to do 7.5ghz+, which I've done on asus but haven't tried on gigabyte. I don't know what their core/module disabling is like compared to asus, and I know their software ocing was less well developed last I tried it on AMD. I am anxious to see what I can make it do. 1090/1100t are the chips to run for this comp, they can score well at reasonable 6ghz overclocks. I have neither. All my good Phenom chips have been sold. Tempted to pick one up for the sake of pride and trying to win this comp, but I could just buy a gigabyte motherboard with that money and not worry about winning.

LN2 has uses I believe in fitting metal components together. Don't know if copper would shrink faster than aluminum, but if you didn't want to chop the aluminum off with a rotary tool, you could try freezing the crap out of the thing and knocking the slug out. The cold would either seize it up worse, or give you just a bit of extra give. Either way, you are pouring LN2 on stuff so its good entertainment.

Extreme GPU cooling on an extreme budget (also has applications for ram or NB cooling): http://www.overclockers.com/extreme-gpu-cooling-extreme-budget

Really nice general usage!

Sata3 SSDs can single handedly surpass 220MB/s XP startup.