Buying Recommendation For Pc Desktop Computer For Chief Architect?


orencons
 Share

Recommended Posts

I would love to see a real world speed test comparison between machines of varying specs.

Download a CA sample plan and do a perspective full overview render of 12 or so passes and then compare the times. All with same render settings of course.

  • Upvote 1
Link to comment
Share on other sites

I would love to see a real world speed test comparison between machines of varying specs.

Download a CA sample plan and do a perspective full overview render of 12 or so passes and then compare the times. All with same render settings of course.

 

 

...or maybe even a RayTrace ;)

 

Graham tried to organize something like this using a

camera view from the Riverstone Plan. Unfortunately 

it fizzled out without getting any significant results.

See the end of this thread. https://chieftalk.chiefarchitect.com/index.php?/topic/4962-processor-core-usage-in-chief/page-9

 

I was kind of hoping it might get some traction but

apparently I was among the minority.

Link to comment
Share on other sites

Graham tried to organize something like this using a

camera view from the Riverstone Plan. Unfortunately 

it fizzled out without getting any significant results.

See the end of this thread. https://chieftalk.chiefarchitect.com/index.php?/topic/4962-processor-core-usage-in-chief/page-9

 

I was kind of hoping it might get some traction but

apparently I was among the minority.

 

The offer still stands. I will run a new series of Raytraces using the Riverstone plan and post them to see if we can get some traction. this time around.

 

Graham

  • Upvote 1
Link to comment
Share on other sites

Not ready to upgrade yet but will as soon as they make a single processor that will out run the dual Xeons in my current system.

 

This will depend upon which generation of chips you choose to compare. There are consumer chips that outperform Xeons in a one on one comparison. The big advantage of the Xeon chips is the ability to have a dual configuration. I am not sure Intel will ever incorporate this feature into their consumer grade series chips as it would potentially undermine the Xeon market.

 

I am very curious as to your comments concerning your system being "brought to it's knees". Given your systems specs I can only assume that this is something that happens when you are multi-tasking while running a Raytrace in the background. If this is the case it is not related to the Xeons power, it is a function of the chips built-in process management algorithms that control interrupts and how requests are channeled through the processor under competing situations. Older generation chips are not as efficient at this as the newer generations, they have done a lot of work on this management side of things in the 6th gen chips.

 

To overcome this issue, especially with high core count chips, you just need to back-off the core assignment in the preferences raytrace settings. just knock it down one core at a time to find the sweet spot. Yes your Raytrace will take a bit longer but with 24 logical cores the impact of losing two or three would not be significant. This is not a work around, it is one of the advantages of having high core count chips. This capability is also built into the Windows OS and you can go into this and assign core allocation to many if not all of your running programs.

 

Graham

Link to comment
Share on other sites

I did a Passmark test on my system not too long ago and the numbers came up just a bit better slower more likely than the fastest i7 at the time so no reason currently to upgrade. That fastest chip was still $1000++ retail so no motivation to spend for a new system to equal what I have now.

 

I have two low speed and low power consuming Xeons (it was the only chips I wanted to afford) and the only way they run well is with the overclock available on this particular (and only) Mother Board. No issues during RayTracing or multi tasking.

 

As far as my system being brought to its knees, I have a couple of Chief models of large houses, not necessarily large files, that I can't work on in 3D, the lag is too severe; 5 - 8 seconds. Probably not a CPU issue but can't rule that out. My 780 video card might be an issue but I'm not spending $500 $650 for the occasional model that slows down.

 

Looking forward to newer, faster chips and the dual Xeon setup has no advantage other than speed, which is a pretty large advantage but the system is huge and complex and really can't wait to get a simpler, fast machine in the future.

Link to comment
Share on other sites

Here's the latest PassMark test on my system for those interested in such things.

 

I don't know what the overall number means and am only focusing on the CPU number, just because I don't know any better. Looks like the CPU number is at 14,679 computer thingys (technical term google it) versus today's fastest i7 at 16,001.

 

My CPU's are overclocked to 3.0 Ghz. the test shows their stock 2.13

 

Interesting, it looks like a new 980 is about 35% faster than my old 780. That's pretty compelling.

 

http://www.cpubenchmark.net/high_end_cpus.html

 

post-302-0-65212800-1457707902_thumb.png

 

post-302-0-92147000-1457707909_thumb.png

 

Here's some more detail. I remember buying a little cheaper, slower memory to keep the system cost down and the test results show that.

 

post-302-0-14592600-1457708704_thumb.png

 

post-302-0-01186000-1457708713_thumb.png

 

post-302-0-12315600-1457708727_thumb.png

 

post-302-0-54646900-1457708736_thumb.png

 

post-302-0-92436700-1457708744_thumb.png

 

post-302-0-26573800-1457708753_thumb.png

Link to comment
Share on other sites

Larry - I think your configuration strategy was good. Using dual Xeons of an older generation at a lower cost per unit can deliver better performance than a new generation single higher priced chip. Your test results demonstrate this.

 

As your Raytracing and multi-tasking is fine it appears that the issue is related to either something within those plans, your graphics card or some odd combination. When you say you can't work on them in 3D, is this lag in both vector and standard view?

 

Graham

Link to comment
Share on other sites

Larry - I think your configuration strategy was good. Using dual Xeons of an older generation at a lower cost per unit can deliver better performance than a new generation single higher priced chip. Your test results demonstrate this.

 

As your Raytracing and multi-tasking is fine it appears that the issue is related to either something within those plans, your graphics card or some odd combination. When you say you can't work on them in 3D, is this lag in both vector and standard view?

 

Graham

Yeah both, Tech suggested it was rebuilding the complex roof (re-building in 3D not re-framing) that was causing the slow down so maybe a faster video card will help but not willing to pay for the upgrade currently. BTW my CPU's are Over Clocked to 3.0 gHZ

Link to comment
Share on other sites

Yeah both, Tech suggested it was rebuilding the complex roof (re-building in 3D not re-framing) that was causing the slow down so maybe a faster video card will help but not willing to pay for the upgrade currently. BTW my CPU's are Over Clocked to 3.0 gHZ

 

Agree - this seems to be a unique situation and it would likely be pointless to spend additional $$$ for this one-off situation.

 

Just out of interest I downloaded the test software you referenced. Ran this and set the baseline configuration to compare all the available dual Xeon L5639's. I only ran the trial version so I can't post the actual results. However if I'm interpreting the results correctly my I7 6700K was right in there with the best of them other than the over-clockers . Seems like the main key is that the 6700K is about 2 times faster on a single thread operation which means that the Xeon needs at least double the threads to match it, add to this that the 6700K runs at a higher clock rate and things start to equal out.

 

Having said this, it does not undermine the power of these Xeons. They are designed primarily for server applications were they can run flat-out 24/7. My chip under a sever application would likely burn-up in a few days. They also are more suitable for over clocking, I could not push my 6700K the equivalent % increase as you are able to do with your Xeon.

 

Graham

Link to comment
Share on other sites

VERY interesting Greg. What is that disk exactly?

 

I wonder if my Samsung SSD needs to be reconfigured somehow? It looks very slow in comparison.

 

Love the clock speed of your Xeon, looks like a nice chip. With two there is probably the added benefit of 24 cores but the performance of your Xeon looks very good.

Link to comment
Share on other sites

Maybe someone with a lot of clout and influence in this group could post a challenge...

- Ray-trace Riverstone plan exterior.

-They should submit the saved camera view and ray trace settings.

Then we could all submit our benchmark times along with a spec list of the machine we used.

Link to comment
Share on other sites

I'm not sure it requires clout but it's been tried so many times here that I'm not sure I'm game.

 

The problem is always the same. Textures, materials, lights, lumens, view, zoom etc. absolutely everything has to match or the test isn't really that interesting.

 

If someone wants to host I'll try and participate but time is limited.

Link to comment
Share on other sites

I'm curious what the actual benefits are with a single Xeon processor, versus a similar throughput i7 CPU, and Chief?

 

Just looking at the Pass Mark chart (not that it's the end all) it looks like the i7 5930 is a touch faster and a touch cheaper than the Xeon 1650.

 

Is the Xeon an inherently better choice? Why? Why not?

 

http://www.cpubenchmark.net/high_end_cpus.html

 

And what's the current fascination with the 6700? It's not that fast in comparison to the faster i7's. Is it SkyLake versus Haswell? The overclockability? For shear throughput why choose a 6700? Budget? Other reasons?

Link to comment
Share on other sites

The attention on the 6700K is mainly due to it's balanced performance. As you can see there are other I7's in the 4th and 5th generations that can outperform it but they are considerably more expensive. Also, some of those performance benefits are not sufficient to warrant those extra $$. Most users would not in the real world notice a 10% or 20% CPU performance difference. If I can output an acceptable Raytrace in say 2 minutes then how much benefit do I derive by only reducing this time down by 12 or 24 seconds.

 

As mentioned the Xeons are designed to run flat-out 24/7 in server applications, that's their primary benefit. A regular I7 is not designed to do this and will overheat and eventually fail under this type of stress, not what you want in a data bank with 100's or 1,000's of servers.

 

Graham

Link to comment
Share on other sites

Maybe someone with a lot of clout and influence in this group could post a challenge...

- Ray-trace Riverstone plan exterior.

-They should submit the saved camera view and ray trace settings.

Then we could all submit our benchmark times along with a spec list of the machine we used.

 

Here is my Riverstone exterior Raytrace - 10 passes, 57 seconds = 5.7 seconds per pass.

 

Graham

post-4793-0-66895800-1457722112_thumb.jpg

Link to comment
Share on other sites

As mentioned the Xeons are designed to run flat-out 24/7 in server applications, that's their primary benefit. A regular I7 is not designed to do this and will overheat and eventually fail under this type of stress, not what you want in a data bank with 100's or 1,000's of servers.

So do you think there's a benefit in Chief? Doesn't sound like but???

Link to comment
Share on other sites

So do you think there's a benefit in Chief? Doesn't sound like but???

 

Yes, especially in the way you configured your system. At the time you got 12 physical cores, 24 logical cores at a great price. Your system even against the latest I7's is still in the top performance range and will likely continue to serve you well for many years to come. The lag you have in some of your plans would likely be the same no matter how much money you through at your CPU's. Keep in mind that the very top performance range is a constantly moving target, they will always come out with something faster, usually in 6 months from ones purchase date. This is now happening to SSD drives, the new PCIe ones currently entering the market makes current SSD's seem like the HDD's they only recently replaced.

 

Graham

 

Graham

Link to comment
Share on other sites

Sorry for not being clearer. I was wondering, given the choice between an i7 and a Xeon (not my cofig but perhaps a newer system) why would one choose one over the other in relation to Chief and how it utilizes the CPU?

 

Greg above has a Xeon chip but it's very rare to see that here in the forum and since Chief doesn't run 24 hours a day what's the benefit, if any? Comparable chips have the same 6 cores and was just curious about the dynamics of the different chips - specifically in Chief.

Link to comment
Share on other sites

I don't believe there is a difference, which ever one gives you the most bang for the dollar. In fact a number of the I7's are actually Xeon chips that do not meet the Xeon stress tests, they just tweak a few settings and label them as an I7. They do this to minimize the rejection rate, not dissimilar to sorting and grading hardwood flooring.

 

Graham 

Link to comment
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
 Share