Rauf

Would like to once again bring up the issue of that all users that have the highest score gets equal points (xtu 742). We have seen lately that benching locked skylakes increases quiet a bit with the result of multiple users get 49hw points. If this continues we will soon have maybe 10 cpus that very easily gives 49 points and that can never be beaten. And then kaby lake comes with perhaps the same principel and hw-points will become mute. Everyone gets 20x49 points...

Thanks for handling it quickly. The one thing I don't agree on is that rules have to be strict, they have to be clear and as simple as possible. Why not limit to xtu freq only, what does cpu-z add? Cpu-z can be tricked...

I have uploaded another run, please remove my previous one and check that everything is correct...these damn rules... You should also check the 5G Skylake tweakers challange. If you have the same benchmark, the rules must be the same! Edit: The site was responding slow last night so I might have accidentally uploaded my first score twice...

Rules say freq determined by xtu! Edit: I have another 1797 run, I'll check if screen is ok for that one and upload as backup if ok. But in my opinion the first should def. be ok.

Of course, I already have. It was mostly meant as a heads up for others.

The rules tab says that xtu don't need screenshot, but at the xtu stage it says it is needed. I'm guessing the later is correct? Also I guess we have to monitor cache freq during run and show it in the screen?

The time depends on many factors. Bigger board will not freeze up as the impact does after some time at full pot. Also if you bench at full pot or higher temps. If you bench cpu only or gpu/mem on cold etc. And of course humidity. Eraser is no point as it makes the board freeze faster. And insulation is not that big an issue if you have some experience and think about what areas will be the ones you need to focus on. Basically you need to think about what gives you the issue and try to adjust next time. A few things you can try while you are running: Bring temps down if you bench at a higher temp. For example -100 is not cold enough so it can cause free water at some spots on the board. Bring the temperature down colder and turn the system off. This may freeze the water and ice do not conduct so it can give you a little more time. Second is to heat up a little if you run a smaller board at full pot and the board is too cold. This will only give you a few more minutes though is it will freeze fast o ce you go colder again.

I have never really used vaseline except for on old vgas. I don't think it is good. I have used LET alot, but I'm actually moving more to just the blue shop paper towels. It is enough for any sort of XOC, and actually better for some applications. I would recommend you to pack these blue paper towels tightly around the memory and the slots. If there is no free air around the memory or slots there can be no condensation. Or get a powerful dehumidifier

it is not overclocking...all I have to say about it

From the little I know about internet security the sites should belong to hwbot since the last dot in the address is the one in hwbot.org From my basic knowledge a fake site would look like hwbot.org.xyzfake

Excellent post from der8auer, the one about 2d vs 3d. I think that from some kind of fairness point of view, 3D should receive the higher points. But points won't fix the 3D decline, or the decline of all XOC. We have previously discussed this and had some ideas on how to make 3D cheaper and thereby more accessible and popular. It is a discussion not for this topic really, but maybe the most important to have. Sadly only a few people seem to care... All money driven sports have already come to the conclusion that too much money hurts the sport. Be it player salaries, money backing, equipment or engineering/development costs. When they get too high and start to influence too much it has to be limited. But I dont agree with the "keep all benchmarks" thinking. We have a point system which is based on the activity of each benchmark. We have to keep the ones we have busy for the points system to be viable. Also there is the aspect of competitiveness. Even now we don't see people with the best cpus/gpus bench all relevant benchmarks because there are too many so they don't need to. Seeking out the categories with the lowest level of competition is a very viable strategy here on hwbot, but the phenomenon is not good for the sport. A high level of competitiveness should be a nobrainer for all sports that wants to thrive.

Haha, waiting for recalculation of league rankings now? Good job!

Very helpful zero... What you need to do is check the top scores in each benchmark and check what OS they are using. The best OS differs from bench to bench. Once you have found out which OS is best, you can either strip them before install using nlite/ntlite, or after install by disabling services. For XP there is a good guide if you google "blackviper xp services"

+1 on the above. But for the efficiency points, this will only lead to people downclocking for screenshot. There are low clock competitions running for this sort of thing also. Considering the rest of the replies, or lack of replies from most of the members, maybe we should scrap the points system? Then we can just bench for competitions or whatever we think is fun. Maybe that's the solution for "benching is not fun anymore"? I would definitely quit the rankings if we got 40 hw point slots. No way I'm grinding like 10 different GPUs every time Intel releases a new generation of CPUs... Maybe benching has turned into too much grinding, there are so many CPUs and GPUs for hw-points, so many benchmarks for globals. I would like to see an "Elite ranking" where we compete only with our top 10 to 15 global scores. And scrap all the outdated benchmarks for globals: do we need two Cinebench? do we need WPrime at all? do we need both superipi 1M and Pifast? Do we need hwbot Prime? Do we need all the 3Dmarks? I agree with Rule that it would be more fun if we compete with quality over quantity. One idea for an "Elite League" would be to have it running in two seasons per year and each season we have 20 predefined benchmarks (4 core XTU, 1xGPU FS etc.) to choose from to make our 10 submissions count.

Nice, please add background as soon as possible as it is vacation time in europe and benching time is limited.

Good Luumi!

What you wrote is actually all lies. Don't spread lies in my name. I hope other people can actually read my posts, and mrscott can continue this discussion in PM if you want to.

As usual you contribute with something irrelevant I think most understand that I meant overclocking as we know it on hwbot. And I don't mean sponsorship to elites etc, but someone has to pay for servers, competitions, development etc.

Thanks for the update! And a good presentation also! First of all, there is an error in the table, but also in the code for the uat server. In the table describing the algorithm, the bonus points go #1, #1, #3, #4, #5. There is no #2, and also, the points for the one that should be #2 is wrong. I guess this is why in the ranking at uat-server, the top 3 benchmark score is listed as nr 2, and sometimes also gets more points than nr 2. Sometimes it is nr4 that is ahead of nr3…strange My main concern with R7 in general is that if we have a points system and a ranking it has to be rewarding. Both for the top players and for the enthusiasts. If you go from a global top 20 to a global top 10 position it has to make a difference. Also if you are an enthusiast and you get a water cooler and improve your score by 1000 points you have to be rewarded. Let’s look at examples: In R6 Fire strike 1xGPU: Going from top 20 to top 10 gets you 7,6 points. In R7 you go up 11 points. It is a little better, but still too little. If we instead consider XTU, going from top 20 to top 10 in R6, you get 10,7 points. In R7, you gain 5,3 points. Does this reflect the effort and achievement? If we look at it from an enthusiast’s point of view it is even worse. Let’s say I have pretty decent rig; I have a 5960X, watercooled and overclocked, and a 980 Ti run at stock. I run the Fire strike benchmark, and get 18500 points = 1 point in R7. If I overclock my GPU and gain 1000 points…I still get 1 point! Might as well kill the enthusiast league… Some other random comments: On the + side: Good to reward 3D-benchmarks slightly higher than 2D. I think the difference looks like a good balance Points scale decently for top 5 positions in R7, compared to only top 3 positions in R6. I would like to see better scaling all the way down to top 10 or possibly even top 20. Good that you can get hw-points for modern hw now. Interesting that WR points scale with relation to top score! On the - side: Some 2D-benchmarks like pifast, wprime, superpi seem to get global points lowered artificially to maximum 89 points? That is too low. WR points are too high. Lower both CPU and GPU points. Or remove CPU WR points all together, it’s not really overclocking after all. Lower GPU WR points to level of CPU points, or even lower. You have to consider that the GP for 4xGPU setups have also increased to 3 times as much as in R6. Way too much focus on hw points in the rankings with 40 point slots. Keep it as it is today, you have already increased max hw points to 53,8. Also consider the industry aspect. Overclocking cannot exist without the vendors. And I don’t think they are interested to see people top the rankings with only submissions on old hw. A thought is that since global 3D points have been boosted, why not balance it by boosting CPU hw points? The top CPU hw points are much harder to get than GPU so it seems fair HW points scaling is too low. It is very easy to get 30+ points. 30 points is a third place in R6… 25 slots for competition points? When we have 25 relevant competitions we can do this, not until then. 10 is fine for now. How are equal scores supposed to be handled? From looking at XTU 2 core it looks a bit random… One final random suggestion: Give global points for mem OC. Even though I don’t like it so much, it is pretty popular, and popular with the sponsors also. Split it into categories like DDR4, DDR3, DDR2 etc.

Thanks!

No need. The board freezes up completely so no moisture anywhere. Cannot bench a whole day, but it is good for around two hours.

You assume people will break the rules and cheat. Yes, easily verifiable rules are optimal, but is the alternative better? People can just as easily cheat with sharing a GPU for 3D-benchmark, why don't we allow that as well? If there is one thing the top dogs know it is which persons have good cpus, so I actually think that CPUs are harder to share than GPUs. I can't see any reason not to do this...

It’s about time, don’t you think? We now have many 3D-benchmarks which are actually 99% CPU-bound. And it is kind of frustrating to see some of the elites submit with multiple accounts using the same golden CPUs, taking up all the high points. With the points-slope being so steep that it only gives good points to the top 3 submissions, hwbot becomes just a battle between Asus and Gigabyte. I love the battle, but keep the submissions to one account please! Examples: Dancop`s 3DMark06 score: 65476 marks with a GeForce GTX 980 Ti Hiwa`s 3DMark06 score: 65315 marks with a GeForce GTX 980 Ti ^^most likely same CPU Dinos22`s Unigine Heaven - Xtreme Preset score: 10523.75 DX11 Marks with a GeForce GTX 1080 TeamAU`s Unigine Heaven - Xtreme Preset score: 10413.28 DX11 Marks with a GeForce GTX 1080 ^^Can’t be sure but probably same CPU Dancop`s Unigine Heaven - Xtreme Preset score: 10448.51 DX11 Marks with a GeForce GTX 1080 http://hwbot.org/submission/3234937_toolius_unigine_heaven___xtreme_preset_2x_geforce_gtx_1080_10299.56_dx11_marks ^^Probably same CPU These are examples, maybe they are not all true, but most likely some are. And I DO NOT say that they are wrong or do not follow the rules. My point is that it isn’t exactly fair and the rules need to be changed so that they reflect the original purpose of the rules; One golden piece of hw should not be used by multiple accounts in the same benchmark. So my idea is just to extend the hw-sharing rule to both CPU and GPU for Aquamark, 3D01, 3D03, 3D05, 3D06, Unigine. Actually, since we have physics tests for the newer 3Dmarks and CPU is not at all unimportant for at least 720P catzilla, why don’t we extend the hw-sharing rule to both CPU and GPU for all 3D-benchmarks?

Thanks guys!

Cool