Adding to Cart…
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.You currently have no notifications.
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.
Comments
I still say it is calculation iterations and the fact the Full core is not needed at the time for that calculation. And as noted before this is not seen ever on my Intel Chips. 100% cores at all times during every render. I have only made a educated guess as to what the AMD is doing. I have never ran anything but a Intel PC. And both my current PC's are Win7 64Bit i7's one a Quad and the other a Hex core. Both will run full on happily until burn out if I pushed them to do so. I see the AMD as doing a better job in that regard not worse.
I truly have zero reason as to why your PC is not maxing the cores fully at all times. I have never pulled the 3Delight engine apart to see (or looked on the source site) what it does. I just render.
Is there another scene or item this happens on, that I might have to hand?
How about the new ready-to-render scene 'sci-fi warrior'? That has a similar looking glowing blade?
Again, no bad feelings Jad, it was a test I neglected to do. I just dropped the old cooler in and ran it. I had doubts.
I don't think the sci-fi things in that free set use ray-trace reflections. I read if an image is used there, it uses the images for the reflection, not the surroundings. I installed the Sci-fi stuff yesterday with the Daz Studio update, haven't looked at anything else yet. I'll give them a try.
The fact that it renders full bore till it gets to that staff (in DontDropTheBall, lol), tells me it is something to do with that staff surface settings or something. that scene rendered in about five minutes without the staff there, just the figure and sci-fi stuff, and volleyball.
Making the staff. It is uterly simple to make the ofending Staff.
Make a primitive cylinder 1in width x 6 ft long
Take the black and white image, and place it in "Diffuse Color" box, and set the background color to Magenta (255,0,128).
Specular Color to 255,255,255 (White)
Take the black and white image, and place it in "Ambient Color" box, and set the background color to ?Lime? (0,255,128).
Take the gray with black stripes image, and place it in "Displacement Strength" box, and set slider to 100%.
min/max displacement can stay at default -0.1 and 0.1 (I used -0.20 and 0.20 to emphasize the glowing grooves)
Set the "Reflection Strength" slider to 100%
Vertical Tiles up to 20
Done, The texture map was more difficult to make then the staff it's self. The 2 inch balls on the ends are the same except 1 vertical tile instead of 20. Set there parent to the staff (Cylinder) after they have been 'Transited' into position.
just place the staff in any scene and watch it fall on it's face.
(edit)
I forgot to put in the all important.
I hereby provide these two texture maps and the instructions of there use for anyone to use as they see fit, FREE of charge.
I re-created the staff:
All cores around 100% throughout!
Ray depth at 16, the staff added about 30 seconds to this 2 minute render:
I just noticed something like that. up close, it uses more CPU. Far away, it falls on it's face.
9m 17s to get to the staff
12m 15s to complete the render.
that's 2m 58s to toil over the staff.
P.S. your renders look good.
(edit)
it takes longer to render a stick, then it dose for hair, lol.
I just noticed something. a sticky-not on my desk for spheres and cylinder primitives, to make them more round.
96 "Segments" instead of 12
and
192 "Sides" instead of 24
There's the extra two minutes of cpu flops time, all them reflective polygons?
should I double the res of the texture files to smooth out the grove edge a bit, or is the raptor teeth edge good?
The toothed edge looks good.
Extra geometry would add to render time because you're using displacement maps.
If you switch the map to the bump channel you should get some speed back.
None of this explains why a core would be idle when it's still required.
What happens if you switch Cool'n'Quiet on'?
I'll check that now. Up close it seams to drop much less often.
Also I increased the depth; Minimum/Maximum displacement to -0.40 and 0.40
I apparently forgot to adjust that on the spheres the first time, they were at -0.10 and 0.10 (Only the cylinder was at -0.20 and 0.20)
O.K. this is on the verge of epic fail. According to the manual, and ASUS's web-sight;
1. Enter CMOS setup screen, and set "Advanced-->CPU Configuration-->AMD Cool'n'Quiet Technology" to [Enabled].
well, everything is now in "Auto" for CPU clocks and voltages (except one setting that only took a min V number, no Auto),
and the only thing on the "Advanced-->CPU Configuration" BIOS page, is a list of CPU stats, and no settings of any kind. Hmmmm.
(edit)
Is this even going to work without the Laptop Battery circuitry on the motherboard? I'm still not seeing percentage numbers under the Processor tree thing?
Ok I'm in 'power saving' mode with 'Active' cooling. I don't think 'Tmonitor' is reading correctly... 3.7GHz!?
Aliens skit flashback.
six meter man.
That's inside this room it's not reading right.
It's reading right man, 5 meters What the hell.
Then your not reading it right.
4 meters... Then one of them looks up at the drop-down ceiling panels, lol.
AMD "Shoulder Grind"-n-"Face Plant" enabled. 23min instead of 12min, it still falls to less then full cores when it gets to the staff.
What are the numbers on Tmonitor? What is the 802.7 top right?
(edit)
now trying a bucket of 32 from the default 16.
A larger bucket size did decrease the idle intensity considerably, now I'm even more baffled.
As for the "Staff Of Render Hell", one last map to finish off the morphing color bands. Leave the Specular Color at 255, 255, 255, and put the blue-ish stripe pic in that box. Everything else is the same as described before.
Good morning,
on the left of Tmonitor is the maximum Turbo frequency 3.7GHz. If you have one or two threads then the chip can put them through at turbo speed (by shutting down other cores).
You get nice red bars on the Bar graph. The conditions for Tutbo are so rare I hardly ever see it.
On the right is the current frequency, at the current multiplier. These chips idle at a multiplier of x4 thats 800 MHz. Mine has a max. of x14 to give me the 2.8GHz Your's should reach x17 to get that 3.4GHz.
Here's Speccy giving the same information in a none graphical way:
Have you looked at the Performance tab in Task Manager to see what it reports the cores are doing? I wouldn't trust some utility built by the processor's manufacturer any more than I would trust benchmark scores by them.
very good point. I don't think Tmonitor is reading correct here, and it could be the socket mismatch. "AM3" CPU in a "AM3+" Socket and motherboard. The socket supports power features the CPU dose not, I don't remember how fine grained the control was (independent core clocks and voltages, or core parts). It was many cups of coffee ago, and stuff I didn't care about being a HPC-cruncher at the time.
In Bill Oddie's words, "Lets have full noise, full speed, let it go". (Truth About Killer Dinosaurs)
In any case Tmonitor is Not showing the CPU usage, just some clock settings somewhere. My core temps alone was telling me nothing is running at 800MHz. socket mismatch?
Yep, Tmonitor is not showing idle CPU usage, just some random clock setting somewhere.
"Wilmap In Motion", Updated "Staff Of Render Hell", new perspective view angle. (SR was 0.5 for this run, Bucket size of 32).
I was supposed to be at my fathers two and a half hours ago, ops. I woke up late. Chat later y'all.
Just a quick wrap up on this, quite some time later.
I figured out, that a "Bucket size" of 32 threw 64 was all that was needed to keep the CPU busy when some chunks of work got stuck crunching reflected ray paths. 3delight has also been updated since then, and I now have a faster newer 8-core computer (My new computer and the newer 3delight is happy between 24 and 32 buckets).
Bucket Size, appears to control two things at the same time. Work-unit block size (or something akin to in number of pixels), and the number of things (Buckets) to run at the same time. It was the second part of that bundled-control-value that was causing my computer to lose traction under demanding buckets of reflection.
Just a minor update to this, tho nothing much has changed. The solution is still to make sure the 'Bucket Size' is sufficient for the scene and computer without excess. mjc1016 made a good point (that I am in complete agreement with) in a Studio beta thread that I will quote here.
So, with my new 8-core CPU (that I've had for a few months), here is the latest run with this Staff Of Render Hell, using Studio 4.8 64bit. Normally I run with the "Bucket size" at 24 instead of 8, and this has not been an issue of late. Some have noted a similar drop in CPU load with beta Studio versions, and increasing the default bucket size has again improved the CPU traction for there renders. So I felt it was worth posting these results with the newer computer, for future reference.
This is the default "ZDG Test Chamber v2.2 UE2 for daz studio" and the "Original" Staff Of Render Hell". The test was to validate the Bucket size vs Drop in CPU load effects with newer computer hardware.