Cinebench R11.5/R15 Rules Revision: Rendered Scene Must Be Visible

A message from the moderating team

TLDR; For all Cinebench benchmarks the rendering scene must be entirely visible in the screenshot and atering benchmark files or the rendering is not allowed

A while ago it came to our attention that there is a modified version of Cinebench around with altered textures which have as main effect to reduce the workload and consequently inflate the benchmark score.

The texture modification can be easily seen when comparing the scene rendered with altered textures with a scene rendered with original textures. Generally altering benchmark files is not acceptable for any benchmark. The so-called "Tex tweak" or "Tex mod pack" resizes the set of cinebench textures from 2928x1810px down up to 4x2px. Over the past days Christian Ney tried many different modifications to the textures files and found the following:

  • The texture file alteration can boost the score from a mere 0.1% up to 200%
  • Most of the time texture modification result in worse performance
  • Modifications that brought more than 1% are easily detectable when comparing rendering scenes
  • Below 1%, it's not evident but it is detectable as well
  • On average texture modification boosts the score by 1-2%

As a consequence of the findings, the moderating staff has update the benchmark rules for Cinebench R11.5 and Cinebench R15 and will proceed with the following steps:

  • Global points will be removed temporarily
  • Screenshots with the rendered scene not entirely visible will no longer be accepted
  • Old scores with the rendered scene not entirely visible will be blocked if they look too efficient for the system
  • Old scores where the rendered scene looks altered will be blocked

The moderating staff will do its best to ensure the integrity of the benchmark leaderboards after the sweep. If you find submissions that the staff may have missed, feel free to report the scores and let us know. We appreciate your help and support!

The HWBOT Staff


79

Ukraine StingerYar says:

What about old netbooks, which have resolution like 1024x600, or 800x480? Rendered scene won't be fully visible there, so is it allowed to submit results "as is", or I have to use external monitor for such situations?

It is difficult to fit 1366x768, so 1280x720 / 1280x768 / 1360x768 as well. Is it okay to submit results like this:

Christian Ney says:

To answer the above question, yes this is ok for mobile submissions. As long as you try to show as much as possible and do not deliberately cover the rendered scene. another example

Additional notes to the news article:
- Not mandatory but highly recommended: Save screenshot as png or use the capture software that is going to be added to the Cinebench rules page. To avoid that a low quality jpg screenshot gets discussed/reported/blocked because it looks like textures were altered.

Also, I would recommend that everyone who has been using altered textures files remove their submissions themselves. From November 1rst, leftovers will be considered an offence and will result in sanction.

No matter the "excuse" ("everyone else is doing it so I do it too", "It's like LOD", "it was not explicitly disallowed", "I didn't know"), I have heard them all, altering benchmark files is not acceptable for any benchmark, period.

Would you replace combat aircrafts in 3DMark03 with pigeons?
Replace the video from the x265 Benchmark with a 4x4px video?
Replace the SuperPi algorithm with a more efficient one?
Change 3DMark's score calculation?
Remove Jane Nash?

Spain gasparspeed says:

Exactly what [MENTION=28623]StingerYar[/MENTION] says. For example I almost always have to bench on a 1280x1024 monitor (those old TFT monitors) since i don't have any better monitor. And I always have to put CPU-Z on top of the rendered image. So in this case, how should I proceed? Still have CPU-Z on top of the rendered image blocking a part of it? Or i would have to use a higher resolution (aka get another monitor)?

Here is a finished R11.5 on my 1280x1024 testbench. I can't fit CPU-Z on the lower margin.

Noxinite says:

:O Cinebench is broken?? I think most of my subs have the image covered up cause it's soo big and there is no space...

Australia zeropluszero says:

Subbed

Russian Federation M1RROR says:

if half of the rendering of the scene is visible, and the other half closed tabs cpu-z it can be? I have old screenshots are, no mods and never was

Czech Republic havli says:

So screenshots like this havli`s Cinebench - R15 score: 290 cb with a Phenom X4 9650 will be no longer accepted?
Well in that case no more cinebench benchmarking for me... :( at least untill I figure out how to fit bigger than 17'' LCD on by table.

United Kingdom borandi says:

Don't need a physically bigger display, just need a higher resolution. Most (all?) laptops have an external video output - hook it up to a 1080p monitor and screenshot it that way.

Christian Ney says:

[MENTION=23093]havli[/MENTION]

1280x1024

Indonesia speed.fastest says:

Any way to do it at 1024x768? Because i dont need graphics driver for Cinebench...

Netherlands willemc700 says:

Christian Ney said:

Would you replace combat aircrafts in 3DMark03 with pigeons?

Remove Jane Nash?

It would be funny to have pigeons and remove Jane Naish.. noway.. maybe here clothes :p
Your right, keep the rules tight there is no room for cheating in competitive sports

Christian Ney says:

[MENTION=20004]speed.fastest[/MENTION]

1024x768

"Acceptable"

United Kingdom ObscureParadox says:

Honestly this is so dumb, any of the experienced guys trying to make excuses should be named and shamed, they clearly knew they were going against the rules with this. Everyone knows you can't adjust benchmark files so why would they think this is any different.

United Kingdom gavbon says:

Should be sanctioned, ashamed and openly named!

United States Splave says:

How about adding pause command lines and affinity priority etc to pifast is that OK? That is also changing the files of the bench. Where do we stop? Maybe such vulnerable to "cheat" benches shouldn't be used here. How about changing the pi-rec.dat2 file for pi? Also cheat?

United States Mr.Scott says:

Splave said: How about adding pause command lines and affinity priority etc to pifast is that OK? That is also changing the files of the bench. Where do we stop? Maybe such vulnerable to "cheat" benches shouldn't be used here. How about changing the pi-rec.dat2 file for pi? Also cheat?


Every bench is vulnerable to cheat brah. There is and always will be a way if one wanted to be tenacious.

Indonesia Lucky_n00b says:

IMHO, I wouldn't be so fast to throw the 'C' word though, since I think sometimes in competitive overclocking, the barrier between 'software optimization' and 'exploiting software loopholes' are not so black and white sometimes. I still remember years ago in my local OC scene, LOD was considered a cheat, not tweak. Also, my memory is a bit blurry about PCMark05, but isn't there a tweak or something that involve changing benchmark files?(cmiiw though, really don't involved a lot with pcmark).

If rules are clearly defined, and the mods team can correctly determine whether a specific run is valid or not, there will be less 'loophole' to be executed.

Now I'm glad the rules are strengthened, so it's more clear which is valid runs and what is not, great job for the moderator team :)

United Kingdom gavbon says:

Lucky_n00b said: IMHO, I wouldn't be so fast to throw the 'C' word though, since I think sometimes in competitive overclocking, the barrier between 'software optimization' and 'exploiting software loopholes' are not so black and white sometimes. I still remember years ago in my local OC scene, LOD was considered a cheat, not tweak. Also, my memory is a bit blurry about PCMark05, but isn't there a tweak or something that involve changing benchmark files?(cmiiw though, really don't involved a lot with pcmark).

If rules are clearly defined, and the mods team can correctly determine whether a specific run is valid or not, there will be less 'loophole' to be executed.

Now I'm glad the rules are strengthened, so it's more clear which is valid runs and what is not, great job for the moderator team :)


If it's just a tweak, why go to lengths to hide it? Surely those at the top of the game should know the difference between a tweak and manipulating files within the benchmark itself

United Kingdom nickolp1974 says:

Firstly thanks to the team in bringing this to the attention of everyone, good job guys but what happens now?? If any submissions deemed to be using this tweek, will it be removed? I dont personally think any further punishment is necessary as there will always be exploitation and this will never be stopped. Would be nice if everyone just ran whatever benchmark at its defaults and we only tunned our hardware, isn't that overclocking?? Competition will always nurture people to bend the rules.

France nvidiaforever2 says:

There,s a problem when u ahve 6400 mhz submissions with an higher score than 6550 mhz ones

Christian Ney says:

You may not recognise the first woman, but you will recognise the second.

NSFW:




Spoiler Alert: It's the same woman

That tweak gave her cancer.

United Kingdom nickolp1974 says:

Why not from this moment on just everyone have a gentlemans agreement to not use any tweaks that alter a benchmark, so no LOD, no cinebench file altering, use a java as set by moderation etc etc. Only your choice of OS and vga driver version allowed. Surely theres brighter people than i can spot anything outside of these rules and then just ban the offending user. Took me 2 years to make that bloody square butterfly to appear!

United States Mr.Scott says:

nickolp1974 said: Why not from this moment on just everyone have a gentlemans agreement to not use any tweaks that alter a benchmark, so no LOD, no cinebench file altering, use a java as set by moderation etc etc. Only your choice of OS and vga driver version allowed. Surely theres brighter people than i can spot anything outside of these rules and then just ban the offending user. Took me 2 years to make that bloody square butterfly to appear!


LOL. If there were more than just a couple gentlemen here, this would never have been a problem to start with.

Greece FireKillerGR says:

nickolp1974 said: Why not from this moment on just everyone have a gentlemans agreement to not use any tweaks that alter a benchmark, so no LOD, no cinebench file altering, use a java as set by moderation etc etc. Only your choice of OS and vga driver version allowed. Surely theres brighter people than i can spot anything outside of these rules and then just ban the offending user. Took me 2 years to make that bloody square butterfly to appear!


ok so after this (fixed drivers, OS, etc. ) + the l&l cooling method the whole oc scene will be now more dependent on luck?

aka what I mean is that this way is an extreme way to deal with the issue.

United Kingdom gavbon says:

nickolp1974 said: Firstly thanks to the team in bringing this to the attention of everyone, good job guys but what happens now?? If any submissions deemed to be using this tweek, will it be removed? I dont personally think any further punishment is necessary as there will always be exploitation and this will never be stopped. Would be nice if everyone just ran whatever benchmark at its defaults and we only tunned our hardware, isn't that overclocking?? Competition will always nurture people to bend the rules.


I agree with you mate, but...

What I would do is the following:

Make an announcement saying people have 2-3 weeks to purge their submissions manually to help clean HWBot up/save time in having HWBot staff spending time fixing this "avoidable" issue.

After this period, go through the submissions and any that are dodgy, remove the submission and ban the user for 6 months for making people waste their time. People who know they have done this, know they have done it...so shouldn't be too hard to spend 5 mins removing a submission

I think it should also be made clear by HWBot what counts as a tweak and what doesn't, but I stand by what I said; people at the top should know better and HWBot shouldn't be scared to do something about those consistently bending the rules.

websmile says:

I don´t see the problem what has to be made clear about benchmarks, what is funny is that people who bench competetively for 10 years or more write bs here
a standardized problem or test that serves as a basis for evaluation or comparison (as of computer system performance)
Standardized means that the integrity of the program has to be guaranteed, if you change files or resolution or whatever that is used to run it, you cheat beause you change the calculation that is done as basis of evaluation period
What needs to be clarified is what tweaks are allowed in general, and hwbot already ruled out in general stuff like speedhacks, wireframe hacks and so on. there is a general staement on this. Tweaks are used to optimize environment a program runs in to make result better, the computation and calculation can not be touched, as well as real time for example to trick the progeram into running faster
You cannot participate in a marathon, enter the subway after 20km, drive two stations by train, re enter at km 25 and claim you won with a new record...

P.S. 30 seconds to use google, find merriam webster will show you the definition I posted above - i hope if helps :)

United Kingdom gavbon says:

websmile said: I don´t see the problem what has to be made clear about benchmarks, what is funny is that people who bench competetively for 10 years or more write bs here
a standardized problem or test that serves as a basis for evaluation or comparison (as of computer system performance)
Standardized means that the integrity of the program has to be guaranteed, if you change files or resolution or whatever that is used to run it, you cheat beause you change the calculation that is done as basis of evaluation period
What needs to be clarified is what tweaks are allowed in general, and hwbot already ruled out in general stuff like speedhacks, wireframe hacks and so on. there is a general staement on this. Tweaks are used to optimize environment a program runs in to make result better, the computation and calculation can not be touched, as well as real time for example to trick the progeram into running faster
You cannot participate in a marathon, enter the subway after 20km, drive two stations by train, re enter at km 25 and claim you won with a new record...

P.S. 30 seconds to use google, find merriam webster will show you the definition I posted above - i hope if helps :)


I don't believe it gets any more definitive than that really! ^^

United Kingdom nickolp1974 says:

Mr.Scott said: LOL. If there were more than just a couple gentlemen here, this would never have been a problem to start with.


Its funny you say this, when i really started to OC properly as it were i found this place so clicky which it still is, i asked about for ages about tweaks etc as i knew there must if been some as i was getting beaten massively by people with similar clocks etc and it was a woman going by the username of Today/oneday whom i beleive was a moderator here and she helped me no end in the tweaking of benchmarks, basically opened a whole new world to me and i am forever grateful to her and this then led me to know HiViZman, another top guy i owe a lot too and i wish he was still around. From my point of view these were the kind of people that drive our hobby forward and are willing to share what they can and help others in there quest.
I know i can criticise at times but i really do not envy the staff at HWBot, it must be such a hard job keeping this place in check and i know personally that there are gentlemen/women among us, its just a shame they get overshadowed by the dark side of OC.

Germany der8auer says:

So we allow LOD on here which clearly makes a huge visual difference - bench 3DMark with LOD 15? You can't see anything anymore and you tell me this is "not altering the benchmark"?
You might not alter the .exe itself but surely it affects the benchmark in exactly the same way as those textures do.
LOD can easily boost your performance over 5 % and more in those benchmarks. The approach is different but the result is the same. In fact LOD is even worse.

I know that other people think different about this and that's absolutely okay. But where is the line between cheat and tweak?

Feel free to argue with me about this but for me a cheat is something like photoshopping results which means you didn't do anything else than producing the result yourself in photoshop.

If you run this tweak and gain 2% more performance, you still ran the same benchmark at the same clock speed for the same amount of time at the same load.
To compare this with websmiles post. This is like running the marathon with different shoes but the same lenght as everybody else.

So if you start about definitions:

Simple Definition of tweak
: to change (something) slightly in order to improve it : to make small adjustments to (something)



Simple Definition of cheat
: to break a rule or law usually to gain an advantage at something
: to take something from (someone) by lying or breaking a rule


It's easy to throw the word "cheat" in the room but I seriously think it's not as simple as it seems.

Christian Ney says:

Little note on LOD: If you ask me LOD should have never been allowed in the first place but well it was. Now it is too late to take it back. Also, not really easy to spot is it? This will change with Time Spy and future 3D benchmarks. I will not argue with the rest.

United Kingdom ObscureParadox says:

der8auer said: So we allow LOD on here which clearly makes a huge visual difference - bench 3DMark with LOD 15? You can't see anything anymore and you tell me this is "not altering the benchmark"?
You might not alter the .exe itself but surely it affects the benchmark in exactly the same way as those textures do.
LOD can easily boost your performance over 5 % and more in those benchmarks. The approach is different but the result is the same. In fact LOD is even worse.

I know that other people think different about this and that's absolutely okay. But where is the line between cheat and tweak?

Feel free to argue with me about this but for me a cheat is something like photoshopping results which means you didn't do anything else than producing the result yourself in photoshop.

If you run this tweak and gain 2% more performance, you still ran the same benchmark at the same clock speed for the same amount of time at the same load.
To compare this with websmiles post. This is like running the marathon with different shoes but the same lenght as everybody else.

So if you start about definitions:





It's easy to throw the word "cheat" in the room but I seriously think it's not as simple as it seems.


I know your point about LOD but honestly, where does it end with the cinebench one? Sure you can adjust it slightly and get 1-2% increase, but you can also change it in a manner which allows you to get 100%+ increase in score. The problem is, where do you draw the line? For me it should be at 0%, render the same image, keep bench as fair as possible and leave it at that.

websmile says:

There is no need to argue, if you alter a program that is used as benchmark this is no tweak. You change the base of calculation. LOD is a tweak that is borderline for me, you can or cannot see it as legit, but it is no direct change of the benchmark program. Your definition of cheat fitsfor altering benchmarks files btw. You leave a common rule that has been agreed on, a benchmark program that is the same for all users as standardized measurement base, by changing it to get an unfair advantage P.S. My main point should be clear, I do not ask for someone banned now over this, or hang people. What should be clear is that people who change benchmark files, or bench non allowed resolution for 3ds, or similar stuff should be banned immediately in the future. Especially because of this I think it is needed to clarify that altering benchmark files can never be allowed

Germany der8auer says:

websmile said: There is no need to argue, if you alter a program that is used as benchmark this is no tweak. You change the base of calculation.
LOD is a tweak that is borderline for me, you can or cannot see it as legit, but it is no direct change of the benchmark program. Your definition of cheat fitsfor altering benchmarks files btw. You leave a common rule that has been agreed on, a benchmark program that is the same for all users as standardized measurement base, by changing it to get an unfair advantage

P.S. My main point should be clear, I do not ask for someone banned now over this, or hang people. What should be clear is that people who change benchmark files, or bench non allowed resolution for 3ds, or similar stuff should be banned immediately in the future. Especially because of this I think it is needed to clarify that altering benchmark files can never be allowed



People adjust pifast, superpi rec dat file, rename .exe of benchmarks to get a higher score. Where is the difference?

Where is that rule written down? I think I checked the rules 50 times. Either I'm too stupid or it doesn't exist and should be added to the general rules.

If there is no rule, you can't break it. It's as borderline as LOD and in the grey zone.


edit:

ObscureParadox said: I know your point about LOD but honestly, where does it end with the cinebench one? Sure you can adjust it slightly and get 1-2% increase, but you can also change it in a manner which allows you to get 100%+ increase in score. The problem is, where do you draw the line? For me it should be at 0%, render the same image, keep bench as fair as possible and leave it at that.


Yea I agree with you. I actually didn't even know that 100% increase is possible until Ney tested it.
This all reminds me of PCMark05 tho where "legit" runs were tweaked 300% over the normal score and nobody bat an eye. Now we see scores which are 1-2 % off and everyone loses their minds - lol

Italy RULE says:

In my opinion there's only one things completely wrong, Dancop could ask to some moderator before post is result, like a lot of people do.The sneaky way make the man guilty no matter why. About tweak/cheat the line is too tiny to be clearly defined, is not the first time a situation like this happen, still haters gonna hate and moderator will moderate. Inviato dal mio SM-A300FU utilizzando Tapatalk

Greece FireKillerGR says:

I will agree with OP on this... LOD has its limitations; modifying the code or the files of a benchmark doesn't... u can really go as far as you want to on this.

websmile says:

Roman, pcmark05 went offline because of this. And on the written rules, I know that common sense and ethics are not estimated high anymore, but when you have to write explicitly that it is disallowed to hack benchmark files, then we can go offline... as said before, the mere definition of the word benchmark disallows physical altering of files because you change the measurement

United Kingdom nickolp1974 says:

der8auer said: People adjust pifast, superpi rec dat file


I knew it!! Care to share?? Priority and thats my lot!

Indonesia Lucky_n00b says:

People adjust pifast, superpi rec dat file, rename .exe of benchmarks to get a higher score. Where is the difference?



Not that I have something useful to add to this discussion, but the longer I read the posts here, the more I'm thinking that I really missed a LOT of these hidden tweaks :D

websmile says:

nickolp1974 said: I knew it!! Care to share?? Priority and thats my lot!


You will be a bit dsiappointed - it is legit imho because you don´t change computation but only allow pause to set affinity and reatime :D

To close this for me, we all decided no one gets banned or sanctioned because we give the advantage of doubt, people might have used this without being aware that this might be cheating. What is important is that we make clear and all agree on the fact that in the future altering benchmark files and thereby changing computation and calculation of benchmarks is no go and will then be sanctioned. Like Rule said, if you have doubts about if something is according to the rules, contact head of moderation :)

TaPaKaH says:

der8auer said: But where is the line between cheat and tweak?
Generally, a benchmark can be seen as a function of system parameters. If you adjust those system parameters (such as clockspeeds, OS settings, driver settings, LOD included) then it's fine. If you adjust the benchmark/function itself, then it's not.
der8auer said: People adjust pifast, superpi rec dat file, rename .exe of benchmarks to get a higher score. Where is the difference?
Following with the above:
- Renaming the .exe file is not cheating since it in no way affects the benchmark process;
- Pre-pausing PiFast with a .bat file is not cheating since it in no way affects the benchmark. You can also do it if you use an external .bat file that calls the hexus_pifast.bat rather than modifying the hexus_pifast.bat itself.
- Modifying the rec.dat file is technically cheating since you are messing with the performance logs that the benchmark keeps for its future use. However, if your system is able to do a 16K in 0.093s or less (and that is easily doable even on C2D) then "using an older version of the log which you might have obtained yourself" is technically not cheating as you are just reverting the system to an older state.
Anything else?
der8auer said: Where is that rule written down? I think I checked the rules 50 times. Either I'm too stupid or it doesn't exist and should be added to the general rules.
I'm afraid to tell you the bad news ... but it has been discussed multiple times in the PCMark era that modifying benchmark files is a strict no-no.
You are a hwbot mod, who else than you should be aware of this?

Germany der8auer says:

I didn't follow all this pcmark05 stuff but if it was already discussed there, why is it still not part of the general rules?

TaPaKaH says:

Hacking the benchmark is already mentioned in the general rules and is awarded with a user ban (part 3.1.B)
HWBOT General Rules and Guidelines
Now we need to discuss what is considered a "hack" as it's not defined anywhere in the rules (not that I can find within 5 minutes, at least).

Personal opinion: modifying benchmark files in any way is "hacking".

Sort of semi-valid argument: there is also nothing in the rules that would currently separate the following alterations to the benchmark
- modifying the render files so that you get 1/2/9000% score boost
- modifying the main .exe file so that you get a WR score on every run
to hacking and non-hacking. So if the first is fine then second should technically be too, right? :)

Sweden Rauf says:

So LOD does not change the benchmark files or the workload? It does, it's just that you do it through a program and don't touch the files yourself. What about mipmap? It's the same thing... They line was drawn at mipmap because it was too extreme. You guys think it is easy to draw a line between what is allowed and not? Then you are not truly creative in finding out tweaks. Here's an example for you: you install a benchmark, let's say a 3dmark in winxp, and then move the folder to win8. Is that a cheat or a tweak? Benchmarks files are "modified"...but they are still the original install files that are allowed in another OS. And consider all gpu driver tweaking in 3d benchmarks. They all affect "the workload", it's not like they actually make your system go "faster". Maybe people would be happy if we were to only allow tweaks listed in the rules of each benchmark... And let he/she who has never affected any workload in a benchmark cast the first stone.

United Kingdom gavbon says:

I think the point here is (correct me if I'm wrong), but Christian Ney pointed out that this tweak can easily net massive point gains. So why then has it only been used to add 2% to a score? Surely if it was a tweak, then you would go max out on it, not make it so it looks like good efficiency.

Sweden Rauf says:

gavbon said: I think the point here is (correct me if I'm wrong), but Christian Ney pointed out that this tweak can easily net massive point gains. So why then has it only been used to add 2% to a score? Surely if it was a tweak, then you would go max out on it, not make it so it looks like good efficiency.

Because then it would be found out. Easy as that, what good is a tweak if everyone has it. Or if it's banned because possible gain is infinite

United Kingdom gavbon says:

Rauf said: Because then it would be found out. Easy as that, what good is a tweak if everyone has it. Or if it's banned because possible gain is infinite


Because it's not tweaking to hide it for other people to use and thus lose the advantage, but it's tweaking so it's undetected by HWBot admins

United States Mr.Scott says:

Rauf said: Because then it would be found out. Easy as that, what good is a tweak if everyone has it. Or if it's banned because possible gain is infinite

The same could be said for a hack/cheat.

says:

is this the benchmark that best differentiates single threaded performance?

Greece FireKillerGR says:

Rauf said: Because then it would be found out. Easy as that, what good is a tweak if everyone has it. Or if it's banned because possible gain is infinite


I think gavbon is right on this one; Pursuing higher % increase would make it look too fishy from the first place, rather showing that there is something that others should find out & use.

Sweden Rauf says:

gavbon said: Because it's not tweaking to hide it for other people to use and thus lose the advantage, but it's tweaking so it's undetected by HWBot admins

It isn't? How many have tried to get the aussies 2xgpu aquamark tweak? Would everyone still be chasing it if it was just 1% gain?

United States Splave says:

So how about all these new junk benchmarks that are affected by the version of java and or using different system info than current highest or even the multiple versions of a bench allowed to used? how about adding obscure command lines to the program file.

Half of hwbot is not "standardized" by any means.
If tweaks are not allowed just give dancop the crown now and we can all save our money and ln2.


People with a criminal history are not usually considered judge material.

Greece FireKillerGR says:

Rauf said: It isn't? How many have tried to get the aussies 2xgpu aquamark tweak? Would everyone still be chasing it if it was just 1% gain?


the aussies tweak afaik is kinda random (whoever has more info pls correct me on this); based on the posts on this thread, this so called tweak gives you the ability to adjust the gain to go for the amazing-unbelievable performance or for the 1-5% advantage that might not be proved/seen.

United Kingdom ObscureParadox says:

Old backup, efficiency through the roof, it's just LOD boys. 64516_zps998fdac2.jpg Photo by ObscureParadox | Photobucket

United States MR. Horse says:

Good grief Another rule pushing higher rez screens, who would have know. I guess its time to find a late model trinitron with higher rez support, I been spoiled by a high end CRT for so long that even the best of IPS and OLED's look like junk. I have a hard time with the idea of down grading for benchmarking.

Germany Flav0r says:

MR. Horse said: Good grief Another rule pushing higher rez screens, who would have know. I guess its time to find a late model trinitron with higher rez support, I been spoiled by a high end CRT for so long that even the best of IPS and OLED's look like junk. I have a hard time with the idea of down grading for benchmarking.
High end CRT with 1024*768? Whats that for an model?

Indonesia speed.fastest says:

For old school i use Sony CPD-E500, resolution can go as high as 2048x1536.

Ukraine RomanLV says:

well, good what not BMP and I can upload such screenshots: http://i.piccy.info/i9/e2f3cfe76eeb983614db936b65fa4c57/1475475460/116534/1055283/2222.jpg ;)

CL3P20 says:

I dont see what all the fuss is about. *This is a legit cheat. (heard of this working manymany years ago..how is it only now hwbot decide to look into this?) *Results using cheats are infractions. *Make a wrapper or remove points for the bench. *Move on. The program/s are easy to manipulate, and have no checksum for rendered scene or any type of device to 'measure' alteration...they only produce score (unlike Pifast and SPi, which do). Wprime FTW, that is all. Thanks.

United States Strong Island says:

..

Romania Alex@ro says:

Splave said: So how about all these new junk benchmarks that are affected by the version of java and or using different system info than current highest or even the multiple versions of a bench allowed to used? how about adding obscure command lines to the program file.

Half of hwbot is not "standardized" by any means.
If tweaks are not allowed just give dancop the crown now and we can all save our money and ln2.


People with a criminal history are not usually considered judge material.


The difference between tweaking java versions or versions of benchmarks is that you test and you are still between "common sense" limits.

What i want to say is that you look for something to improve performance,you will hit lower scores,better ones and finally settle for one scenario which gives you consistent better scores.

With this Cinebench adjustement you are looking for a way to improve performance but be carefull not be "too much" of an improvement while still make things looks like lega e.g not altering the image too much.

It is exactly the same if u wanted to photoshop a 3dmarkscore and edit subtest FPS then u use 3dmark calculator to generate a valid final score to photoshop that one too.

Australia Sgt Bilko says:

I don't think this will affect my old subs that much but I will have to remember this going forward....

United States Splave says:

Compressing a texture is no different then LOD, only its not a cheat when the registry handles it.

Arent you glad this has come to light? I posted the how to cheat XTU and nothing was done, Ive already found a way to make the screen look correct and mess with the bench files and nothing will be done either besides removing the "OLD tweak" results.

Maxon doesnt care, we dont make them money. 99% of users dont even know Cinema4d is the program that control c4d files / cinebench.

Remove r15 and r11, they both work the same way.

Sweden Rauf says:

Agree with splave, there is no going back for CB. No big loss, it's just brute force anyway. There are better benchmarks out there...

Indonesia Lucky_n00b says:

Not that I have a certain attachment to Cinebench, and certainly there's better benchmark out there(which scales better with memory for example), I just think that this benchmark does something which easily resembles 'real-life application', compared to those like Geekbench or HWBOT Prime.

Personally, I find both single-threaded test and multi-threaded test in Cinebench quite easy to relate to when comparing certain CPU and architecture for normal users(minus that it is not so memory-heavy, so it put less stress to memory bandwidth).

If there's nothing more that could be done to verify result then I'm okay with it being taken away, but would it be possible to make a 'wrapper' that check the integrity of its files? Or would that be too much work?

United Kingdom gavbon says:

If you removed half the benchmarks with LOD issues and those that could be manipulated...how many would actually be left?

Greece [MSI] OGS says:

Lucky_n00b said: Not that I have a certain attachment to Cinebench, and certainly there's better benchmark out there(which scales better with memory for example), I just think that this benchmark does something which easily resembles 'real-life application', compared to those like Geekbench or HWBOT Prime.

Personally, I find both single-threaded test and multi-threaded test in Cinebench quite easy to relate to when comparing certain CPU and architecture for normal users(minus that it is not so memory-heavy, so it put less stress to memory bandwidth).

If there's nothing more that could be done to verify result then I'm okay with it being taken away, but would it be possible to make a 'wrapper' that check the integrity of its files? Or would that be too much work?


Maybe something like that yes; wrapper that takes screenshot and also has file checksum comparison (if that would work)

Christian Ney says:

To answer [MENTION=26407]gavbon[/MENTION], if we were to remove all hackable benchmarks, the benchmark list would be empty. Now some have better security/integrity than others.

United States Gunslinger says:

This kind of stuff makes me glad that I'm software illiterate and don't ever have to worry about "tweaking" my way into trouble.

Australia zeropluszero says:

Likewise

Australia macsbeach98 says:

Wrapper would be the way to go I dont think getting rid of it is a good idea variety in benchmarks can only be Good

says:

Allen ( Splave ) said it all.
Why spend our hard earned $$$ for a pre-sold game ?

This whole thing folks ( cheat or tweak ), reminds me of top of the line
athletes, whose medical teams continuously search for doping drugs which will pass the exams without being recognized.

I feel so tired and naïve.

BY THE WAY - what is LOD ???

Australia ozzie says:

at the end of the day george, whether its the athletes, medical teams or cheaters in the computer benching area, all theyre doing is cheating themselves, and causing unnecessary drama and inconvenience for others moderating and competing, its a very selfish attitude and with which shows they havent got what it takes to do it the right way for their best times/scores and performance from their mind, body or hardware

and LOD as i understand it means Level Of Detail, like in 3D benches, in rivatuner for example the LOD can be adjusted for less or more detail on the actual screen in how it looks, im not fully sure on how it works fully , im not experienced enough, im sure there is someone here that can explain it much better than i

says:

ozzie said: at the end of the day george, whether its the athletes, medical teams or cheaters in the computer benching area, all theyre doing is cheating themselves, and causing unnecessary drama and inconvenience for others moderating and competing, its a very selfish attitude and with which shows they havent got what it takes to do it the right way for their best times/scores and performance from their mind, body or hardware and LOD as i understand it means Level Of Detail, like in 3D benches, in rivatuner for example the LOD can be adjusted for less or more detail on the actual screen in how it looks, im not fully sure on how it works fully , im not experienced enough, im sure there is someone here that can explain it much better than i
Well said my friend - My question about LOD was pure sarcasm :D Egoism - the epitome of the human substance. Tricks, cheats, tweaks - FOR WHAT ? For a higher place in the Global Rankings ? Or to justify the money and gifts that a handful of "players" receive here in exchange for their "valuable services" to vendors ? I was not aware that hacking computer software ( benchmark files ), was part of the "abilities" someone must have in order to be competitive.

Bulgaria I.nfraR.ed says:

And yet again whoever did this will get away with it, because he "didn't know it is not allowed". Common, especially top dogs know exactly what they are doing. I don't even know the whole story, but these kind of things demotivated me to bench for hwbot. Haven't done it since quite long time now. Better spend my hard-earned money in other hobbies (like mountain-biking), where you get rewarded for your actual skills. You get the point.

says:

I.nfraR.ed said: And yet again whoever did this will get away with it, because he "didn't know it is not allowed". Common, especially top dogs know exactly what they are doing. I don't even know the whole story, but these kind of things demotivated me to bench for hwbot. Haven't done it since quite long time now. Better spend my hard-earned money in other hobbies (like mountain-biking), where you get rewarded for your actual skills. You get the point.
Exactly sir. We are on the same page. It is really meaningless. Let me tell you all how all this will finally end. 10-20 "players" competing between them who is the best tweaker-cheater.

United States MR. Horse says:

wile I do not quite know how this cheat woks, but whats to stop someone from say using steganography? EDIT: I guess the score would give it away, but that would make things even harder.

Christian Ney says:

[MENTION=74887]liqmet[/MENTION] You can start checking out both Cinebench rankings

Australia zeropluszero says:

Why would you encourage this

United States coolhandluke41 says:

nothing will change as long as you won't hold ppl accountable for their actions..adm. included

Please log in or register to comment.