Lucid Virtu MVP: Revolution in benching?

[rbuass]

Lucis is for gamers an not for overclockers - it will kill the fun

[Trouffman]

Here is what i think :

 

* Lucid Virtu / MVP can run on any iGPU / GPU they want, and will support multiple GPU (did i say so ?)

* The reason it's tied to intel chipset only and some of them only is... LICENSING ! No point on having to pay XMilions $ more to lucid to get that on AMD platform or previous because... it make no sense on selling more cheaper stuff by paying more someone else and this will arm your profit.

 

* My opinion is that it should be illegal or at least red flagged and detected for a reason : with the software version (virtu / MVP) it act with the same concept as physix and other illegal tweak... IT DOESN'T render everything OR it select which is better to render, so we slip from having a benchmarks for testing hardware to a benchmarks for User experience.

 

Benchmarks goal is to measure the performance of a product or a system. Hwbot focus on gathering the data to compare hardware between them and system in a matter of hardware based solution.

 

Lucid product implies on the User experience and the optimization of redundant work.

 

-> that's why as Massman kinda stated, regular benchmarks cannot be used to measure impact of lucid logix and game responsivness, etc that was the case for ages in the reviews site discussions... what benchmarks make senses etc...

 

 

And for Lucid...

* They have hardware and software solution, for software it should be illegal or be a special categorie as it is a really interesting feature for daily use. As they are Hooking the real intended work of a GPU.

 

And honestly... to detect it... it's either hooking th hook or checking if some specific process / code / call are made

 

More details soon enought..

[ME4ME]

I'll compensate for Dinos postive vote. Im greatly against using this MVP stuff.

 

For gaming? .. sure, if it actually gives same quality graphics.

But for competitive benchmarking, no. It messes up to much, and kills GPU overclocking in some extent.

I see no reason why we would want to use it.

 

I vote: no.

[Mr.Scott]

For gaming? .. sure, if it actually gives same quality graphics.

But for competitive benchmarking, no. It messes up to much, and kills GPU overclocking in some extent.

I see no reason why we would want to use it.

 

I vote: no.

Seconded.

[Massman]

Without knowing the details, or tried the software myself.. one problem occurs to me.

If HWbot is to accept MVP, then SB-E, AMD Phenom, Bulldozer and future high-end CPU's will all be useless for 3D benchmarking since they have no IGP.

 

Ehr, what?

 

Following that logic, we should've never accepted on-die cache. Or a CPU with an integrated memory controller (no more NB!). Or a CPU with multiple cores. Or ...

[der8auer]

This decision is realy hard and I also don't know which is the right way. The easiest way is, like you said, to follow the rules of FM even tho we didn't do it in the past.

 

We should decide which would be the best for HWBot - regardless on our own opinion.

 

MVP on HWBot:

 

Advantages:

- Reach the highest scores possible

- Be up-to-date with the newest technology

- ?

 

Disadvantages:

- Results could be highly affected by updates of the software (next update gives 10% more boost -> bench everything once again...)

- Limited support for different GPUs

- Unsure development in the future (Lucid Hydra!)

- Comparability will be affected due to different scaling on benchmarks and different cards

- Multi GPU not supported (1x GPU will be even faster than 2x GPU)

- Result affected by refresh-rate. -> just use 55Hz CRT and your result is much higher than 60 Hz LCDs

- Benchmarks get highly CPU limited (just bin a lot of CPUs and keep your card on air to win) srsly

- ?

 

Did I forget something?

 

selfquote ftw

 

Still looking for a real reason why we should even add this here on HWBot?

 

IMO there is 0 advantage for HWBot.

[knopflerbruce]

Well, there's no gain in allowing LOD, and a bunch of other tweaks either.

 

A bunch of your points are also not that good, for example, 3/4-way 01 scores are slower than 2-way + singlecard. That was never an argument against 01.

 

Limited support os also not that relevant, as it will work on a bunch of different generations. no support for G80 etc is irrelevant, it's the same as saying that like 3DMark11 is shit because it wont run on ancient GPUs at all. I don't see why we have to know that this software will be updated for generation after generation. if this software is available, then it is, and if they won't update it after 2 years or something, then so be it. It's not a big deal. Then we just don't have MVP subs for those generations. CPu dependancy is also not that relevant, this should be about principles, not if some people don't like that the GPU is less important for 3D if this software is allowed. Would it be OK if this software made all benchmarks 100% GPU dependant? We also have refresh rate dependant benchmark today, although not as dependant as n this case. Still, I dont see the issue as long as the refresh rate isnt stupid (like 10Hz). 55Hz is fair, and IMO anything at 25Hz and above.

[der8auer]

Yea but we can't see which refresh rate was used. So just use some tools and force the refreshrate to e.g. 12 Hz. Will look crappy as fk (which is exactly not the thing MVP is supposed to be) but you get like 400% higher scores than with normal 60Hz LCDs.

 

Once again: What is the advantage to have MVP here on HWBot? I asked this like 5 times no but no one seems to have an answer.

[Hondacity]

Its not a feature per se, its an improvement like a new generation cpu from intel.

 

What advantage did sandy bridge give to hwbot when it came out? of course better scores. everybody had to rebench including 3d01

 

Sli hack on x48....what did that do?

 

Windows vista/7 lets us do more than 2 video cards....

 

Basically a step forward ... for hwbot.

Edited March 20, 2012 by Hondacity

[knopflerbruce]

That's kinda my answer too. We don't benefit from much, only thing I can think of is new, popular benchmarks (= more hits), so if you can't find a gain for HWbot, it's quite normal:p

[ME4ME]

Ehr, what?

 

Following that logic, we should've never accepted on-die cache. Or a CPU with an integrated memory controller (no more NB!). Or a CPU with multiple cores. Or ...

 

You totally missed the point, or maybe just ignored it.

 

Either way, i think there is alot more negative to come by allowing MVP then positive. I really see no reason to use MVP here at Hwbot. All we want is benching and being able to compare score with others. Don't make it futher complicated and ad software that will be basically necessary to achieve a competitive score. Keep it simple, and let the hardware do the work.

[der8auer]

Basically a step forward ... for hwbot.

 

Normally I think the same way: new tech = step forward. But this time I think it would be a step backwards!

 

We don't always have to say yes to every new technique - same as we did with physx. If we don't profit from it at HWBot we don't need it. Simple as that.

 

The only "advantage" you see is that we have a new technique. And even you say it's not a real advantage. But we have about 10 realy strong disadvantages. So where is the point adding it?

 

 

That's kinda my answer too. We don't benefit from much, only thing I can think of is new, popular benchmarks (= more hits), so if you can't find a gain for HWbot, it's quite normal:p

 

Well you get 10 more hits due to MVP but lose 100 members who have enough of benching everything again because the whole hwbot rankings will be crushed because of MVP - and I will be one of them.

 

If you add a public poll what the community thinks about MVP you will get a clear result: no

 

 

Either way, i think there is alot more negative to come by allowing MVP then positive. I really see no reason to use MVP here at Hwbot. All we want is benching and being able to compare score with others. Don't make it futher complicated and ad software that will be basically necessary to achieve a competitive score. Keep it simple, and let the hardware do the work.

 

^^^^ strongly this

Edited March 20, 2012 by der8auer

[I.M.O.G.]

I don't see why we have to know that this software will be updated for generation after generation. if this software is available, then it is, and if they won't update it after 2 years or something, then so be it. It's not a big deal. Then we just don't have MVP subs for those generations.

 

If its permitted and allowed now, then loses support after a year or two... I don't have a point in this question, I'm just trying to figure out what this could mean. But if it loses support when will the records set on MVP be broken? I wonder if it could do weird things to the global points, where in 3 or 4 years we have much more powerful hardware, but to compete for any top 3D globals we're still falling back to older MVP hardware. Dunno if that makes a difference, however I like the motivation to run newer and more expensive hardware to see what it can do... If new more powerful stuff is less rewarding to benchmark due to MVP, that could be a negative to some people I guess.

 

I don't have a firm position one way or another. In post #158 I shared some other thoughts I had, and rereading it I guess it sounds like I support MVP in a separate rankings. For similar reasons to why Futuremark is looking into detecting if MVP is active I suppose.

[knopflerbruce]

If its permitted and allowed now, then loses support after a year or two... I don't have a point in this question, I'm just trying to figure out what this could mean. But if it loses support when will the records set on MVP be broken? I wonder if it could do weird things to the global points, where in 3 or 4 years we have much more powerful hardware, but to compete for any top 3D globals we're still falling back to older MVP hardware. Dunno if that makes a difference, however I like the motivation to run newer and more expensive hardware to see what it can do... If new more powerful stuff is less rewarding to benchmark due to MVP, that could be a negative to some people I guess.

 

I don't have a firm position one way or another. In post #158 I shared some other thoughts I had, and rereading it I guess it sounds like I support MVP in a separate rankings. For similar reasons to why Futuremark is looking into detecting if MVP is active I suppose.

 

I don't think the benefit will so big that you'd have to wait 3-4 generations to get better scores without MVP. Maybe the MVP "bonus" equals about one generation. That's my guess so far.

 

We shouldn't think of all those if's and but's, what if FM is bankrupt in June? Then no more support for any 3DMark/PCMark version. Is that a valid reason not to add new benchmarks?

 

*************************

 

I feel alot of the reasons people post in this thread are either irrelevant or have been ignored in the past for other benchmarks. it would be much better if we just said "OK, we have no good reason to not allow it, but it's what the users want" instead of trying to find excuses that make no sense at all:D

[dinos22]

Impact is a lot more significant than one gen change

[Gunslinger]

How many more pages before we can get a poll?

[dinos22]

so you dont want to do the testing in the community before you decide, that makes zero sense!

[George_o/c]

Some initials tests by Stelaras using a 9600GT.

 

Post #7, #10 & #17 http://www.hwbox.gr/news-vga/16081-lucidlogix-mvp-ti-einai-pos-mas-epireazei-poia-i-stasi-ton-futuremark-hwbot.html

[Gunslinger]

so you dont want to do the testing in the community before you decide, that makes zero sense!

 

Yeah, I'd rather clock my GPU or CPU higher to get a better score than have some software trick the bench into thinking it ran faster than it really did, you can call me old fashioned I guess.

[Hondacity]

Thanks George o/c and Stelaras( retired? are you retired? )

 

The gains are stupidly crazy high.... LOOOOOOOOOOL

[Trouffman]

Once again... MVP is hooking DX call, it trick the bench / game to not calculate everything and especially duplicated calcul...

so it does affect who render what and how. And that will evolve with the software update... same issue as VGA drivers.

 

The software should not be allowed as it modify the integrity of the benchmarks, similar as pHysix that use the GPU on a supposedly CPU only bench. The hardware version that is a on real card should be.

 

This is a great tool and evolution but that need TESTING, TESTING and TESTING... and as our greek friends shows... benchmarls are completly tricked by the software as the method to calculate, render etc give HUGE improvment in scores.

 

 

(NB : there is a more serious issue about this... how to be sure the frame as been calculated and not dropped to re-used one... as MVp is kinda hooking it.. it would be like hooking the hook until another hook of thehook appear to hide it...)

[oanvoanc]

http://www.hwbot-mvp.org

 

MVP would skew the whole hwbot thing. not like a new CPU/Gpu generation but much much more.

so why not just create a new league/callitwhatever with MVP enabled to separate "normal" scores and "ssoftware thweaked" scores. just like a playground for people who want to play with it.

[Massman]

Ok, some good news. We can perfectly detect whether or not MVP was used for a Heaven benchmark run.

[Massman]

You totally missed the point, or maybe just ignored it.

 

Either way, i think there is alot more negative to come by allowing MVP then positive. I really see no reason to use MVP here at Hwbot. All we want is benching and being able to compare score with others. Don't make it futher complicated and ad software that will be basically necessary to achieve a competitive score. Keep it simple, and let the hardware do the work.

 

I missed the point then. I re-read and still miss the point ...

[George_o/c]

Thanks George o/c and Stelaras( retired? are you retired? )

 

The gains are stupidly crazy high.... LOOOOOOOOOOL

 

hehe he has that user title just for fun If he benches that much now that he "retired", what would happen if he was fully active?