egm_xt

I was surprised as I kept increasing the freq. and it just kept going. It is an SL8K4. There is this comment in cpu world about it " I have this cpu bought as tray model ,it states on cpu 3.40ghz/1mb/800fsb, sl8k4 philippines , it is one of the latest made by intel for 478 socket with northwood characteristics but prescott capabilities , thats why it has g1 stepping & the motherboards need bios updates to accept it." Actually thinking about the comment in cpu world that I have just read while searching in the web after I looked at the SL number, it does seem odd. Northwoods are 130nm and Prescotts are 90nm. CPU-Z identifies it as Prescott and 90nm. The fact is... it does clock very well (This was on ambient air - around 24C and the cpu got to about 52C using VERY high flow fan). I had this CPU for a few years in another board. Putting it in the P4C800 made a big difference.

Valeu! Thank you NoM$_YesLinux! Still have to Vmod the board and may be try a few more things. You are doing a great job on the 754 stage )

Valeu... Thank you. Did not have time (and dice) to optimize the pifast run But it was fun ...

Is it allowed to use Semprons (462) with Thoroughbred or Barton cores?

This is great! Athlons are really fun. I do have a few questions: 1. What is exactly core 21,22 (Barton, Thoroughbred, Thornton or something different...)? 2. I guess this would also allow Athlon XP and XP-M, is that right? 3. I want to participate and ask exactly how to get the invitation? 4. What was the reason for this month competition be different requiring the invitation unlike the other Challenges? Thanks

Thank you 1BadMoJoe for not telling my boss. My other office computer is a FIC VA-503 with a K6-2 /533 and is quiet for now

Congratulations Placid. Very well done!

Thanks Placid. I did try to add ice but oddly it did not work and actually I could only achieve lower frequencies.

Thanks to all, but as we can see the competition is only starting...

Fantastic job!

Thank you. I believe it may improve cooling the chipset better. It was VERY HOT with the standard little fan.

It looks to me as a much better solution. This way the motivation for benching stuff that has not been benched or with just a few results is back. I voted Yes. To address the less points in the 11 to 19 range a sugestion for fine tuning would be: one table from 1 to 10 participants and another for between 10 to 19 like this: ------ 1 to 10 ----- 11 to 19 ---- 20+ rev3.0 1 ------- 2 ----------- 3 ------------- 3.6 2 ------- 1.5 --------- 2 ------------- 2.6 3 ------- 1 ----------- 1.5------------ 1.9 4 ------- 0.8 --------- 1.2------------ 1.4 5 ------- 0.6 --------- 1.0 ----------- 1.1 6 ------- 0.5 --------- 0.7 ----------- 0.8 7 ------- 0.4 --------- 0.5 ----------- 0.6 8 ------- 0.3 --------- 0.4 ----------- 0.4 9 ------- 0.2 --------- 0.2 ----------- 0.2 10 ------ 0.1 --------- 0.1 ----------- 0.1 I don´t believe this is a serious problem but if it is not hard to implement I believe it could help balance things. Thank you.

Thanks for taking care of this issue so quickly! Actually while I was having a hard time to write the post about it you were working on it

I have to acknowledge that it takes courage to make such a drastic change. As it is, and at first contact, I personally preffered much more the old version (rev.2) and would opt for smaller changes addressing each issue at a time: number of GPU/CPU, unpopular/popular hardware, achievements changes, etc... As people say: If it ain´t broke don´t fix it. Or as we say in Brazil about soccer teams: ("Nao se mexe em time que está ganhando") something like: "You do not change if the team is winning". (after all nobody can ever know for sure all the reasons why this is happening... ). Sometimes changes are needed but the consequences are hard to foresee. Please remember that to have winners and 1st places we need loosers and 2nd places. Since it is all relative depending on how high you set the bar there will be more or less people willing to take part. There is no way to tell what is going to happen in the long run but please understand that it is normal for people to feel bad if they were playing with one set of rules and it suffers a big change - specially with backwards effects reducing (sometimes very hard earned) points. But as you say there are bugs to be fixed and to be more practical I will add some numbers to one bug I belive has already been spotted and really left me with a strange feeling. For me it is a very serious one based on the fact that it can change significantly the scores of many people who have less points - lets say less than 300 - and these are the vast majority of the people at Hwbot and the ones that allow the first ones to have significance in their ranking. - results for cpuz with 5 participants (Cyrix M-II 300) 1st 1.8 points 2nd 0.3 points 3rd 0.2 points 4th 0.1 points 5th 0.1 points More than 50% received more than 0.1. The 1st received 6 times more than 2nd (1.8 to 0.3) and this looks strange - results for cpuz with 31 participants (Pentium 3 celeron 900) 1st 5.8 points 2nd 4.5 points 3rd 3.3 points 4th 2.5 points 5th 1.8 points 6th 1.3 points 7th 1.1 points 8th 1.0 points 9th 0.9 points 10th 0.7 points 11th 0.1 points 12th 0.1 points . . 31st 0.1 points less than 30% received more than 0.1 points (the new "zero") the 1st to second ratio is 1.29 (instead of 6 like above) but the drop from the 10th to the 11th is from 0.7 to 0.1 - 7x, also strange Quickly looking around I found this to be consistent. For very few results we are close to 50% awarded and there is too much difference between 1st and 2nd (totaly out of proportion with the difference between 2nd to 3rd). For more results the difference between 1st and 2nd looks "normal" but the difference between the last score with more than 0.1 and the first 0.1 score is now out of expected proportion. I may have looked at a few bugged results but it does not seem so. Anyway this was a huge undertaking and I hope that with small changes from now on you can make rev.3 as fun to play or better than rev.2. Have a nice 2010

Ticket ID: 705 Priority: Low I have submitted some results for K6-2+/533. As far as I know these are the first results for this processor (I attached a link to the picture of it). It was accepted and the points correctly counted but when you try to see the processor specification and hall of fame you get the K6-2/533 instead of the + version. Thanks.

Thank you for the reply. I somehow thought that when I first submited a result for this processor there was only one option but was not sure. Anyway I was just not very happy with the questioning of the submissions if there was only one place to put the results just a few days ago. Anyway that forced me to make a few more rounds in the bench and almost a full day fighting with that BE. But now I have the 2 sets of measurements. Thank you.

I got no response or opinions but I believe this is something that needs some time. I hope the moderators will look at this issue and find a solution. Yesterday I was thinking about the 3 options I could see and it seems to me that merging the results is not an option at all. So I can only see 2 solutions: Move all non Black Editions 5000+ Brisbane to the right place and the value of the scores in this place would considerably increase with a decrease in the value of the 5000+ BE scores. This would be the perfect solution although I am afraid it is not really something that can be done easily. Leave it like it is with probably more than half of the processors in the BE place not being a BE. It should be no major problem as long as everybody understands that a non BE version of the same speed and the same core can be posted in the BE section (I am talking about the specific Athlon 64 X2 5000+ case). Anyway I got a bit worried with this and want to solve it. So I could get a 5000+ X2 BE processor and runned it in my setup. I have to say that it is not as good for over as the non BE 5000+ that I have. Clearly this is just a slight variation that may happen in all cores (some more than others). This BE is not being able to reach in my setup the same top frequencies as the original 5000+. Still after quite a bit of work I did manage to get some results and decided to post them and move my 5000+ scores to the Athlon 64 5000+ X2 Brisbane place, even knowing that probably more than half of the processors in the BE section are non BE. For me this case is solved but I do hope the moderators look at this case and may be find an even better solution than I could see since this fact was used to question my submissions.

I have posted results on the Athlon 64 5000+ X2 BE that are from an Athlon 64 5000+ X2 Brisbane. Why? This is the place where almost ALL 5000+ X2 Brisbane are. The problem was that the wprime32 and wprime1024 results were the best and this caused some problem. Just a quick count shows that 11 of the first 20 results of wprime1024 for 5000+ BE are from 5000+ Brisbane. Also at least 8 of the first 20 results for wprime32 also are from 5000+ Brisbane. Considering that the BE is a Brisbane core with unlocked multiplier this could make sense and that was my reasoning. Anyway there are 82 results listed for 5000+ X2 BE and 1 result listed for the 5000+ X2 Brisbane ( there are a few more but this is what is listed in the processors page). Therefore the number of points are totally skewed because of this practice of placing results from 5000+ Brisbane in the 5000+ BE section. How to solve this? 3 ways: all the 5000+ Brisbane should be removed to the proper section or they should be merged or just leave it like it is considering that there are only disadvantages of placing a 5000+ Brisbane on the 5000+ BE list. I would like this to be addressed. Thanks