Jump to content
HWBOT Community Forums

Different types of overclocking leagues!


Recommended Posts

Hey boys and girls,

 

Given the recent commotion about the Overclockers League being totally unfair (:P), we have tried to come up with a different style of overclocking league. As a try-out, we've made a very rough sketch of what could be an AMD overclockers leauge. The link is not visible through the front-page (since it's not an official feature), but you can check out what it's all about here: http://hwbot.org/rankings/amdleague/

 

As you can see, without too much issues we can make rankings based on specific hardware or manufacturers. This page is a nice concept of one of the ideas we've been playing with to seperate the extreme overclockers from the regular overclockers. Instead of thing top-down (eg: pro league), we've made an alternative going bottom-up. This means that we do not seperate the super-seeded from the regular-seeded, but the extreme from the non extreme.

 

The enthusiast league:

 

- cooling limited to ambient (= stock, air, water)

- frequency of each CPU and VGA limited to undisclosed value to prevent cheating and golden samples ruining the game

- no seperate result ranking based on cooling, points are gathered from the normal rankings(*)

- possibly: exclude expensive hardware from rankings

 

The extreme league:

 

- no cooling limitation

- no frequency limitation

- points are gathered from normal rankings (*)

 

(*) the normal rankings are rankings based on the hardware/manufacturer limitation. So, for this AMD league, the points are not calculate based on the entire database, but only based on the results with AMD/ATI hardware. This means it's possible that certain benchmarks (eg: 01 and 1M) have less weight than the normal rankings because they are a lot less popular than when you include intel hardware.

 

The idea behind using the same rankings for the enthusiast and extreme rankings is that there will be noticable transition when you switch from ambient cooling to extreme cooling. To show you how it works, this example:

 

- my best score on ambient cooling is rankined 89th in the superpi 1M ranking; I get 20 points for that.

- those 20 points are added to my total for the enthusiast ranking; I now have 180 points

 

- the next day, I use LN2 to cool my cpu and get a score that is ranked 15th in the 1M ranking; i get 60 points for that.

- these 60 points are added to my total for the extreme league

 

Since only your best submission is ranked, the submission with ambient cooling now receives 0 points. So, the total for my enthusiast league has dropped with 20 points. This method applies to all rankings and categories, which makes it a necessity to only choose for ambient, or only choose for extreme overclocking (since you cannot have two accounts in hwbot). So, you can be the Extreme King, or the Enthusiast King ... but not both. This, to make sure the extreme overclockers play in their own league.

 

Note: the current ranking has not been calculated as explained above. It would take too much time to do it properly, and time is something we do not have available currently. At this moment, we just make a sum of all points that are added to AMD/ATI scores.

 

amd-league.png

Link to comment
Share on other sites

So I take it global points from say both wprime x1 and x3 are added to the amd league score?

 

Going to have to make cooling selection obligatory as it seems to of screwed with the enthusiast rank already.

 

The current implementation is just: Points = global + hardware (from AMD/ATI only scores)

 

Ideally, we'd have the time to re-build the HWBOINT engine so it can calculate the rankings of the AMD/ATI scores seperatelly and give points based on that. So #1 of AMD would get full points.

 

In fact, in this design, you loose points if you submit a better score with Intel or Nvidia :D

 

Cooling choice will obviously made 'obligatory' (= not filled in equals no points).

Link to comment
Share on other sites

Hey massman,

 

We had a competition in our forum last month... fixed clock, aircooling only... it fails.

Because of very good and expensive prices, the users went crazy.

Some users benched at higher frequencies and decreases the clocks for the screen.

Some people benched under chilled water or some subzero methods, other cooled down the air with LN2.

 

So which clock limit is normal and extreme?

Is a single stage extreme? For some user of course... but for LN2 benchers not really...

What about the tempertures, is there a limit in the normal section?

 

I think it can work, but it will be very hard so check the results.

 

How can you check the cooling method?

 

Or I didn't understand this new feature... then I need to read the first post again :D

 

Thanks for your hard work :)

Link to comment
Share on other sites

The issue is indeed that we can't check the temperature. There are two things this concept has:

 

- We impose an invisible MHz limitation to each HW that is based on the average overclockability on air/water. This way, we can remove the entries with 7GHz 980X "on air" as well as exclude golden samples on air cooling

 

- We put the air/water scores in the same ranking as the extreme scores (like it is now). This has the advantage that there will be no direct competition in the air/water league, but everyone will have to push their setup harder anyway to get more points. It's not as simple as just beating your direct competitor. Also, this has the advantage that people who use LN2 regularly will not be able to both appear in the LN2 and the air/water ranking.

 

Our definition of extreme is "everything below zero".

Link to comment
Share on other sites

Downclocking for screenies will be a problem, as discussed every time we wish the world was full of honest people

 

Chillers to hold at +2, +3 are still a few hundred £, but buying hardware to be top of your league is ok if people are interested in doing it :)

 

 

Obligatory: boo-hoo dont **** with the points system again, its unfair to at least 4 people :P

Link to comment
Share on other sites

Ah I see...

 

So a very good sample of a CPU/VGA, even if it was luck to find a golden one, will be not noticed in the normal section, because the clocks are higher then the average overclockability? That is sad to hear...

 

With my own words, just to check if I understand all things right:

The enthusiast and extreme league are just two more seperations in the rankings, like single or dual GPU. The best result (most poinst) either in enthusiast or in extreme league counts to the profile, but not both.

 

So it is possible to go up to the top of overclockers league just with watercooling?

Because there are more air/water results then sub0 results (I think so)... when the number of results in every league influences the points, like now, air/water cooling will storm the overclockers league :)

Link to comment
Share on other sites

So a very good sample of a CPU/VGA, even if it was luck to find a golden one, will be not noticed in the normal section, because the clocks are higher then the average overclockability? That is sad to hear...

 

One of the problems of the current overclocking is that it's dominated by golden samples ... even if it was just by luck, but in most cases because huge amounts of money is thrown at it. For the enthusiast league, we want to keep things low-profile and not let golden samples be ruling the league.

 

With my own words, just to check if I understand all things right:

The enthusiast and extreme league are just two more seperations in the rankings, like single or dual GPU. The best result (most poinst) either in enthusiast or in extreme league counts to the profile, but not both.

 

So it is possible to go up to the top of overclockers league just with watercooling?

Because there are more air/water results then sub0 results (I think so)... when the number of results in every league influences the points, like now, air/water cooling will storm the overclockers league :)

 

This kind of league is different from the Overclockers League we have now and will also be treated as such. It's a side project ... no worries :D

Link to comment
Share on other sites

What about those of us who run extreme cpu cooling but non-extreme gpu cooling?

Unless I'm missing something, we get hosed.

Admittedly I can't get especially pissed about that considering that this looks like it has a chance of making air/water and normal retail hardware productive.

Link to comment
Share on other sites

CPU with LN2, but air VGAs will go in the Xtreme part.

 

As mentioned in the first post, this concept is working from the bottom up: we don't try to separate the super-seeded overclockers from the 'normal' extreme overclockers, but we separate the casual overclockers from the extreme.

 

Also, this is just a concept ... nothing has been put in practice yet.

Link to comment
Share on other sites

Well, what sort of CPU clock limitations should we use? IMO it should be fairly difficult to get above it on air/water, otherwise it would be a pure tweaking contest (or downlocking contest, if you wish...) and even though it's fun, is it what we want? For Wolfdale I'd sya 5GHz is the lowest limit we should have, alot of good chips can do that. Maybe 4.5GHz for Nehalem/Gulftown, or is that too low?

Link to comment
Share on other sites

Have seen plenty of wolfdales above 5.0 on water. That's just your average E0 imo. I would raise the bar to say 5.4 benchable, much higher you are telling porkies :P

 

Currently on AMD I would say anything much over 4.7 benchable on normal ambients atm is kind of out of the realm...

 

Maybe to set these limits you should use some well trusted members of the community to test a few normal retails on normal air. Max spi 1m = max clock speed :)

________

Lovely Wendie99

Edited by El Gappo
Link to comment
Share on other sites

The biggest issue I see is that of turning down clocks to the acceptable limit after running a bench, it won't be that difficult to find the hidden mhz limit in the 2d benches, spi1m and wp32 especially.

It looks to me like another situation like that which exists with pifast right now, the only way to detect a cheat is to look at how it compares to other scores, which in turn means that if the cheat is kept low (say 1-5%, just enough for a gold) it's essentially undetectable.

 

Only thing I can think of would be a CPUz wrapper like your Heaven wrapper, it runs and monitors CPUz (and/or gpuz) and the benchmark, and then reports back with what it saw.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...