New Engine works with Hyperthreading?

2 Questions re Cubase 8 Performance,

(1) I have a quad core i7 with hyperthreading (8 cores). Does Cubase still have an issue with hyperthreading?

In this Steinberg article: http://www.steinberg.net/nc/en/support/knowledgebase_new/show_details/kb_show/optimizing-windows-for-daws.html , it mentions to turn off hyperthreading (which I can not do in BIOS). Is this still the case with C8?

(2) Will I get a better VST Performance to CPU performance from Cubase 8, than Cubase 7.5.
Have a look at the attached, Cubase is at 90% where CPU is at 20%. Any solutions to help with this appreciated.

Gigabyte X58 motherboard UD5 (July 2009) / Intel Core i7 2.66GHz Quad Core (8HT) / OCZ Gold Series PC3 12800 1600MHz 8-8-8 RAM (2Gig x 6 = 12 gig) / Dual-OS: Window 7 Home Premium 64bit & Windows XP / Lynx AES16 PCI Card connected to a Mytek 8x192 ADDA Soundcard

No Steinberg software are supporting HT directly**, but it doesn’t/shouldn’t have issues with HT either. In this day and age :wink:
But HT may benefit other “crap” running on your computer.

New computers and software have no problem running on a HT enabled pc. Since a couple of years, I have yet to see any trouble.

PS. That Knowledge paper is waaaaay old, and is not valid anymore IMO.

**Discussed in lenghts with an old Steinberg insider (Thanks Chris B), we ended up buying the same computer without HT.
This computers were of corse DAW computers only, running Cubase/Nuendo exclusively.

+1. That “daw advice” knowledge page causes more harm than good.

Thanks, you seem knowledgeable. I appreciate that!

Do you have any thoughts on what could be my problem re point (2). More detailed post here,

http://www.steinberg.net/forums/viewtopic.php?f=181&t=70253&p=406572#p406572

I’ve reached the limit of my knowledge to sort, thanks, j

Regarding your second question, you have to remember that the MS CPU meter and the Cubase ASIO meter is two entirely different beasts.

The ASIO meter has much more going on (what to meassure) than the CPU usage alone.
In short, the ASIO meter in Cubase meassures all things audio related, every real time processing, plugins etc you throw at Cubase, in addition to pure CPU usage.
So the MS CPU meter will always show differently (lower) than from Cubase’s ASIO meter.

To sum up, the CPU resources used is only a portion of the ASIO meter readout (not to be compared).

The most important thing is if Cubase 8 is more efficient than Cubase 7? Usually a newer version needs a bit more resources, but… At this point I will say, it depends. But it looks very promising IMO.
Remember, this is/can also be, very computer dependent (no two pc are alike… well :wink: )

I have not compared scientifically, but at my place Cubase 8 feels a bit snappier (on both Win7 x64 and Win8.1 x64) and in direct “project comparison” (C7.5 vs C8) Cubase 8 seems to use a bit less resources.
And in situations that uses the benefits of the new ASIO Guard 2, C8 uses pretty significant less resources. Important as your project grows.

When opening a relatively heavy loaded project, I can now run at a much lower buffer settings with the ability track through the software at low latency.
The ASIO Guard 2 is waaaay better than how the first version ASIO Guard were implemented in C7 (that was a hit and/or miss) :mrgreen:

Read up on the ASIO Guard feature.
Again, very short: You’ll have a Low input buffer and a High playback buffer (this may be so “in short” that I’m waiting an attack, or two :wink: )

All I got time for now (late saturday evening where I live).
I Hope it answered your question. At least enough to keep you going for C8 :slight_smile:

I have kept HT off in older versions and it helped performance. In 7.5 I turned it back on and it gave me a performance boost of around 20% less ASIO. Cubase 8 seems to get the same performance boost from turning HT on. Individual mileage may vary…

I do not seem to be getting a performance boost from upgrading from 7.5 to 8 - but I believe that I may in the minority. I monitor at 96 samples and did not notice an improvement for either my smaller projects or the larger ones vs. 7.5. This is something that I am still checking into. I do like the concept of how ASIO Guard plays back tracks with a longer latency and only focuses short latency on tracks being recorded / played live. I believe that for very large projects that this will help.

John

Just a note to my previous post - if I turn off “Record Enable” for my current track I do see a performance improvement. This is due to to the whole project being played back at a longer latency through ASIO Guard - as designed. This feature is very convenient vs. having to change your ASIO settings in Device Manger.

John

Same here, I have HT enabled since Nuendo 4 and also with Cubase 8 I have large benefits with HT enabled

Yes, on modern pc’s and software HT should be kept enabled.

But WHY Steinberg do not remove that waaaay to old knowledge paper, is beyond me :confused:

Please Steinberg, remove that ancient Knowledge paper of yours. It does more harm than good in these day and age.

+10000

100% correct! HT should be ON the last 10 years.

Best Regards
Freddie

Or in other words: don’t buy a software these days, which isn’t truly optimized for multi core threading and! hyperthreading. But it’s up to company and marketing to clearify this and communicate the benefits.

Cubase and Nuendo and all other modern plugin and softwares support multi threading and Hyperthreading.

No worries.


Best Regards
Freddie

i keep randomly going back to my UEFI and turn HT on for heavier projects, because i keep hearing here how it should be on-- and I have yet to see a performance improvement. In every single case HT has rendered an almost-maxing-out project unplayable when HT had been enabled (while keeping the same latency; 6core i7).

According to a Steinberg insider (former employee) NO Steinberg software uses HT, but supports and uses multithreading of corse (split on x nr of physical cores).
But it should, and do, work well with HT enabled on your computer. So…

No worries anyhow.

That’s because NO Steinberg software uses HT.

This I have from a lengthy discussion with a former Steinberg tech employee, who explained this (while he were still at SB) in detail.
In a DAW only pc scenario he were clear on that you could get the same result with an i5k processor, and we ended up putting together the same pc.

Important!
The above discussion were about a DAW only computer.
This may differ for a pc used for “everything” else besides a DAW. If you are a gamer or video editor it absolutely makes sense to go HT.

PS. Turning on/off HT can mess with underlaying tasks and processes, if you have much going on on your computer.

I stick with the former Steinberg employee on this matter. At least til Steinberg tells us otherwise.
But Steinberg are good at letting positive myths going on and on… and on :wink:

As I have said before, if Steinberg good tell us things like this in the open, this forum would be a much friendlier place. Much “fuzz” could be avoided IMO.

Peace out :slight_smile:

wow, that’s cold.

I know :smiley:

PS. Just my attempt to provoke Steinberg to enlighten us (in case things have changed) :wink:

Thanks. Very clear explanation, it helps.

One further question: Is there anything I can do to improve the ASIO meter read out? Bar increasing buffer size? On my current computer.

& What I’m also aiming to figure out is: Would getting a new computer help?
ie, If I had a new computer, i’d get the latest new powerful CPU & Motherboard, would I be able to have get more VST’s on a project? Or, does there come a limit where regardless of the power of the computer, Audio can works within certain confinements & that’s more due to Cubase itself working with Audio at 44.1 with various VST’s (and CPU power doesn’t come into it, after a certain power available point has been reached)?

In my instance, only 20% of CPU is being used, though Cubase has maxed out. If I had a new motherboard & CPU, it may showing 10% of CPU being used, but still Cubase is maxed out, cause it’s Audio working with Cubase at & it can it’s own internal limitations regardless of the computer’s abilities.

Similar example, I just upgraded to 2 x SSD’s (one for OS & one for my Cubase Audio files). It made no difference on VST Performance on this project :cry: though, I appreciate the quick load times! :smiley: // So, this upgrade made no difference to audio, unfortunately… though, a worthy improvement never-the-less

Thoughts?

James

Gigabyte X58 motherboard UD5 (July 2009) / Intel Core i7 2.66GHz Quad Core (8HT) / OCZ Gold Series PC3 12800 1600MHz 8-8-8 RAM (2Gig x 6 = 12 gig) / Dual-OS: Window 7 Home Premium 64bit & Windows XP / Lynx AES16 PCI Card connected to a Mytek 8x192 ADDA Soundcard / Cubase 4 on XP, Cubase 7.5 on W7

James - I read your other post. Couple of thoughts / questions:

  • it looks as though I have the same / near same mb as you - and it has HT and the other settings mentioned in the Steinberg article. Worth checking again. I get pretty good results from my setup.

  • you should post you system specs in the signature - this will help others in suggesting ideas to help you.

  • could you provide the layout of the project for the posted picture - number of VSTs, tracks, etc.

  • maybe a stretch - but try trashing your preferences to see if that helps. sometimes an errant settings has caused ASIO issues with my setup.

John

Just to add a recent test on the subject as of 2018 and Cubase 9.5 Pro; I just ran a few project loads and exports both with and without HT, using an i7-6700K @4.6GHz machine.

There is no difference in any load/performance times, and export times are the same within tenths of seconds, ie.no discernible difference whatsoever in any use that I could notice. Used rather a heavy project that took 2.32 minutes to load and 2.27 minutes to export, and a much smaller one that did both in well under a minute.

The temps and fan speeds were unaffected as well.

The only interesting difference was that looking at performance graphs, with all 8 logical cores on, CPU usage was reported at around 50% all the time, with just one core going at about 100% and the rest 7 just around 20% or so.

However with just 4 cores, all cores reported about 100% usage as well as overall it was around 100% constantly.

Nevertheless this had no effect on actual performance speed, temperatures, or anything that I could see.