[Half Life]

Hey MaxTroll, you don't get to call the shots here as to what gets discussed and how -- why don't you go stink up somebody else's forum with your crap?

[brodie_geers]

OK I did some testing with the Benchwell scene on my system (see signature), and here are the results for SL15 :

HT ON :
5m13s - Benchmark 2048

HT OFF :
6m54s - Benchmark 1550

So the difference is less then I thought, but significant anyway.

Thanks for that. I overlooked your post when I posted an earlier comment. This would actually be a performance increase of 20% instead of 30 which others state. That´s within the band with of my original statement of 15% on average for modern renderers. For software I´m involved with I can´t quote figures, but I went through my notes and I compared it with public available benchmarks and I would still say it´s around 15% with Mr at the same level as Maxwell (20%) and most Max bundled renderers apart from Mr at the lower side. It´s of course very scene and hardware dependent. Bottlenecks occur depending on this combination. A silly example: if you run out of memory due to scene and texture size you will get met more effect of HT. However in every normal benchmark this doesn´t happen. However anecdotal results from people with sub par or badly configured systems could also lead to higher effective HT speed.

We´ll have to wait for Dmeyer for a more detailed report about his renderfarm tests.

Max

all these sidelines about Cuda and HT should be in a thread about HW and Maxwell instead of in the real time preview topic. However it developed this way. Maybe it´s best to just let it go and open a new topic about the preview the moment it´s released. Just a suggestion to the mods.

Just so we're on the same page here...

It depends on whether your calculating the time or the benchmark but neither seems to be 20%. The time results in a decrease of 24.4%. The benchmark results in an increase of 32%.

Time: 313s/414s x 100 = 75.6% (75.6% - 100 = -24.4%)

Benchmark: 2048/1550 x 100 = 132% (132% - 100 = +32%)

-Brodie

[max3d]

OK I did some testing with the Benchwell scene on my system (see signature), and here are the results for SL15 :

HT ON :
5m13s - Benchmark 2048

HT OFF :
6m54s - Benchmark 1550

So the difference is less then I thought, but significant anyway.

Thanks for that. I overlooked your post when I posted an earlier comment. This would actually be a performance increase of 20% instead of 30 which others state. That´s within the band with of my original statement of 15% on average for modern renderers. For software I´m involved with I can´t quote figures, but I went through my notes and I compared it with public available benchmarks and I would still say it´s around 15% with Mr at the same level as Maxwell (20%) and most Max bundled renderers apart from Mr at the lower side. It´s of course very scene and hardware dependent. Bottlenecks occur depending on this combination. A silly example: if you run out of memory due to scene and texture size you will get met more effect of HT. However in every normal benchmark this doesn´t happen. However anecdotal results from people with sub par or badly configured systems could also lead to higher effective HT speed.

We´ll have to wait for Dmeyer for a more detailed report about his renderfarm tests.

Max

all these sidelines about Cuda and HT should be in a thread about HW and Maxwell instead of in the real time preview topic. However it developed this way. Maybe it´s best to just let it go and open a new topic about the preview the moment it´s released. Just a suggestion to the mods.

Just so we're on the same page here...

It depends on whether your calculating the time or the benchmark but neither seems to be 20%. The time results in a decrease of 24.4%. The benchmark results in an increase of 32%.

Time: 313s/414s x 100 = 75.6% (75.6% - 100 = -24.4%)

Benchmark: 2048/1550 x 100 = 132% (132% - 100 = +32%)

-Brodie

Time as that is an universal and comparable stick to measure performance gains so this specific example is indeed 24.5%. I estimated it and used an approximate figure as like I said it will depend on scene and configuration, but you´re right that as a ball park figure I should have said 25%.

So my original statement about 12 + 12 HT cores being 14 instead of 24 should be 16 based on this single measurement. Based on DMeyers it would be close to 24 cores, based on comparable renderers I would expect it to flatten out over multiple tests on 12 - 13 but I´m open for more empirical tests.

Max.
Max.

[max3d]

Hey MaxTroll, you don't get to call the shots here as to what gets discussed and how -- why don't you go stink up somebody else's forum with your crap?

Normally I don´t even react on these kind of posts as they are better ignored. I do however wonder what puts you in this position. I mentioned before that I was registered years before you were. Contributed in the very early days with money (bought 16 licenses which I finally had to return because of the RC1-RC5 debacle, but still that money was used to finance development) and with technical posts about several subjects. I had access to very early builds sometimes and tested problem cases. I developed scenes to pinpoint these problems and all that was without any compensation or whatever. I was just an early adapter of the Maxwell way and tried to help out.
I didn´t post the most damaging tests as they were confidential but they were available to the alpha testers to help progress instead of criticizing from the side lines.

Even your specific interest: materials, I spend quite some time on complex IOR materials to help people understand how they worked. I don´t know what you contributed but I can´t remember me even using the kind of words you use.

I don´t hope you´re a representant of the new generation of posters. This used to be a community of like minded people who were driven by a desire to get Maxwell to the top of unbiased, physically correct renderers. Early adapters who spend their time and sometimes even money to help NL get there. People like Mihai, Mike Verta, Tom, Thomas Ann and numerous others who often became alpha team members but already worked hard before they were formally asked. Yest they were sometimes critical but luckily NL noticed their contributions and embraced their knowledge as either artists or technology experts. There were a few very black months around the RC1 period where people who contributed a lot got banned or just left themselves, but a lot returned the moment NL became wise and opened up.

So I hate flame wars and don´t want to participate in it as it´s a waste of time and sets the wrong tone on every forum, so I don´t expect you to answer. I only hope other posters and readers either ignore all of this because it doesn´t interest them or profit from the contributions people in this topic make. At least that has been the spirit for a very long time. Don´t forget that Maxwell was the new kid on the block back in 2005 and nobody used it in his production flow. They were just open for new developments and were aware that a good Maxwell would be the future of Arch/Viz rendering. If the community changed just let me know and I will keep silent and change my signature line.

Max.

[Half Life]

2005 was a long time ago and you haven't contributed much of value since you've been back other than criticisms of everything including how this forum is moderated and who should post what where. I'm not impressed.

You don't rule this forum or this thread for that matter -- you want to contribute? then contribute... if you are just here to pimp your agenda, then get lost.

History lessons are pointless, you've moved on from Maxwell and have another agenda for being here -- which serves as nothing more than a distraction from the people here who actually want to share and learn about how to use Maxwell better.

[tom]

Hey, everybody cool down please.

[Bubbaloo]

Maxwell users are passionate!

[rusteberg]

i like peanut butter sandwiches...

[dmeyer]

i like peanut butter sandwiches...

no jelly?

[rusteberg]

i like peanut butter sandwiches...

no jelly?

sure, i'll splurge every once in a while

although, i prefer a simple fold over accompanied with either a fresh cup of black coffee, or ice cold milk!

[Half Life]

I like natural peanut butter with honey -- but I'm kind of a freak, I stir it up stick it in the fridge and eat it straight out of the jar.

Best,
Jason.

[max3d]

Hey, everybody cool down please.

Tell that to my failing air conditioning! It´s hot in southern spain. How is Greece now?

[Richard]

I've noted from a post on the PushPullBar that yesterday Thea has launched their realtime preview feature. I hope NL aren't long off on their release!

[rickyinmotion]

no more news about maxwell realtime ?

[JTB]

no more news about maxwell realtime ?

Good, someone still remembers how this topic started!

Some news would be more than interesting!