Benchmark

[PFABC]

[PFABC]

Jay,

 

quote:

---

Keith, you bring up the machining of STL as you feel that mastercam does not do well with STL. do you feel this is a large issue?

---

Actually it is not STL machining that I was referencing as much as progressive stock models creation and usage for remachine operations. Yes this is an area I feel that Mastercam is weak in. Just this last week I was doing just as described and the operation kept losing the assocative CAD stock model. The generated stock model was still where it always had been. Try it in X3, create a HST roughing operation, verify it, create a new STL file from that now verified model, copy the roughing operation changing it to a remachining operation, change the cutter, parameters and finally go to drive surfaces and select the STL CAD file you created.

 

Process and verify that and repeat the rest for a second Remachine operation....and some where after that your first Remachine op will go dirty....on closer inspection of the selected geometery in the op you will find that the original CAD file is now missing from selection.

 

Now this is not a really that big of a deal as I can reselect the STL, the point is a one of stability and that this is just one more thing that doesn't work the way it should on top of having to re-process the path. I just don't have these kind of issues with other software/s that I use, they are faster at processing, more stable, better work flow....I won't go on.

 

quote:

---

I know that all have used the one file, it was just a base of one file that all share to see how the computers or as show here hardware doe's make a difference. I am sure that there are more files besides the one that came as a sample. but does this really set much of a difference on how the hardware with compare to other

---

No that's not what I'm am say at all I'm just saying maybe its time to throw in some other variables not just for the sake of processing speed but to run the software thru alternate senarios. Think of it as the difference between a drag race (similar to what is being done in the benchmark today) and the Indy or Grand Prix (Which dictates that is not just about speed, not saying that speed doesn't count)

[NeilJ]

quote:

---

Now this is not a really that big of a deal as I can reselect the STL, the point is a one of stability and that this is just one more thing that doesn't work the way it should on top of having to re-process the path. I just don't have these kind of issues with other software/s that I use, they are faster at processing, more stable, better work flow....I won't go on.

---

Good. It belongs on a blog. A blog I'm looking forward to reading and telling others about if it's well done.

[Thee Rickster ™]

PFABC

 

quote:

---

real world file

---

Yeah the bench mark is one thing to time by....

 

I have countless files where 'one' path take 20 to 30 mins to crunch. I bring the customer data in and it may have 3 to 6000 surfaces upon importing

 

Waterline/rest material and add cuts as well as pencil cut paths take for ever.(no difference in cimatron or surfcam either)stability wise, I would have to agree with you, MC has to be fine tuned by the developers thats for sure

 

I usually create the path get it starting to crunch and hit 'esc' and keep the path to crunch later. I have mold base paths that take sometimes hours to crunch doing them all at once.

 

So I crunch them overnight because they take so long.

 

m2c's

[PFABC]

God forbid a revision the morning after

[sandro]

quote:

---

quote:Hi

3:53

 

Doesn't count. Your in Switzerland. You have metric clocks and most likely running metric Mastercam. Just kidding about the clock but not about MC metric. We run the inch version and there is definately an unfair time advantage for metric. [band Head Against Wall]

---

Dear bogusmill

Even in Switzerland, the minutes have 60 seconds.

3:53 in Switzerland are 3:53 in Usa or somewhere else.

And if i open the bench-file, MC will switch to inch.

Why shoud i kidding, you're angry because your Xeon is slower or what?

[Guest CNC Apps Guy 1]

quote:

---

...Now that these times have been achieved I'm wondering if anyone is open to trying it on a real world file that will separate the men from the boys...

---

[bogusmill]

quote:

---

Why shoud i kidding, you're angry because your Xeon is slower or what?

---

Or what sandro. The what being happy.

 

The first Core 2 duo machine I built had a 1.86Ghz processor OC'ed to 3.4Ghz and it took the benchmark lead and held it for 2-3 months. Slowly others passed me then a guy with a Dell machine and a slower CPU,(dual quad Xeon's) and memory took the lead by 45 seconds.

 

Then I got a rare opportunity to build another PC for me so I decided to try and reclaim the lead cheap. After a little research of this thread I found that Xeon's seemed a little faster than thier C2D counterpart's & quad core's have twice the L2 cache of dual core's. My computer guru told me about Xeon's different cacheing method and after the build I confirmed that with a comparison of a Q9650. So I reclaimed the lead again for 2-3 months and now 3 have passed me. It's inevitable but seeing why I was passed gives more clues what MC likes. It seems to like Vista x64 with 4 or 8gigs of RAM and more CPU speed never hurts.

 

I currently have Vista x64 loaded (5 sec. gain) will be getting 8Gigs of RAM in the next few days, and will sell my Xeon X3350 for a X3360 that has a small increase in multiplier. There is also a BIOS revision to install on the Rampage. After all I'm not made of money and this is my home PC so thats as far as I can go and if that doesn't retake the lead I'll have to move to Switzerland where the best watches are made, a minute is 60 seconds long, and metric is faster than inches.

[Matthew Hajicek - Singularity]

quote:

---

...Now that these times have been achieved I'm wondering if anyone is open to trying it on a real world file that will separate the men from the boys...

---

Just to put things into perspective, when Roger and I first ran this file as a benchmark (back in 2000 or 2001) it took about two hours. Roger had built a computer for home using a P4 800 with Rambus memory, and I had put together a box for running demos using an AMD 800. I forget what the exact times were but the AMD was about twice as fast.

 

Now that this file is taking less than four minutes on the cutting edge systems, maybe it is time to put together a more variegated test file. Roger?

[Jaz]

quote:

---

maybe it is time to put together a more variegated test file

---

I agree. Maybe something that would take 20-30 minutes on the current fastest box. I would definitely like to see something with some biggish 5-axis toolpaths. If we do come up with something maybe a 'New Benchmark' thread would be appropriate.

[sandro]

I'm sorry if i don't understand your words in a correct context.

English is not my first language;)

 

1.86 oced to 3.4GHz, that's realy great.

 

I don't think that MC on Vista will be faster.

Too many unneeded processes are running on this OS.

Xeon or C2D, i don't no wich Cpu is performing better with the same core speed.

But, this is the right place to find it out.

 

And btw, the fast computer is also my home machine.

The office-computer is much slower. (HP XW4400, E6700)

[Thee Rickster ™]

quote:

---

God forbid a revision the morning after

---

So far that has only happened to me once bout 10 yrs ago with different software.

 

engineering changes usually only happen in the casting business after the customer sees the parts and fit and functions etc. weeks later... [phew]

[bogusmill]

quote:

---

Maybe something that would take 20-30 minutes on the current fastest box. I would definitely like to see something with some biggish 5-axis toolpaths. If we do come up with something maybe a 'New Benchmark' thread would be appropriate.

---

New benchmark is needed, maybe several. Not everyone has multiaxis. I'm thinking 3 benchmarks and of course get them included in MCX4 CD package.

 

1. A mold or wing. 3 axis + surfacing - Basic

2. 5 or more axis cutting + high speed - Advanced

3. Graphics only, takes minutes to load and chokes when rotated - Graphics

 

quote:

---

I don't think that MC on Vista will be faster.

Too many unneeded processes are running on this OS.

---

Vista does have too much running but then so does XP vs. Win2K. Airbus here is running Win2K x64. Vista x64 is x64 first x32 second, XP is x32 but later made to run as x64. Vista is the better 64bit and Vista can handle 128Gb of RAM, XP is more like 16Gb.

[Matthew Hajicek - Singularity]

Maybe a verification benchmark too. I'm looking forward to the time when I can rotate an STL compare view without having to turn it off , rotate, then turn it on again (or zoom out a bunch).

[NeilJ]

Mastercam In SolidWorks is being billed as having:

 

"Expanded support for cutting STL files"

 

See:

 

http://www.mastercam.com/CamZone/NewsZone/Article6.aspx

 

My guess is that this is because SolidWorks uses Techsoft3D HOOPS.

 

I would also guess the rest of solid cut part verification in Mastercam In SolidWorks will be better than stand alone Mastercam also because of SolidWorks use of Techsoft3D HOOPS.

[bogusmill]

Any of you that are thinking of building a new machine might want to read about this one. This CPU-Z speaks for itself!

 

http://www.hardwarelogic.com/news/132/ARTI...2008-10-20.html

 

 

quote:

---

My jaw dropped as I found myself hovering over 600 FSB with little difficulty. After some time of finding a stable OC, I settled on a blazing 4.8GHz overclock via the 600 FSB and 8x mulitplier.

---

 

[ 10-22-2008, 01:19 AM: Message edited by: bogusmill ]

[Tony35]

14 min 43 sec

pentium 4 cpu 2.60 ghz 1.50 gb ram

programs open when run

mcx3 open 2x

solidworks

emastercam

music folder

i need a new computer

[Bruce Kanzelmeyer]

Bogusmill,

 

Do you have any idea what kind of cooling they had? Must be watercooling. On my overclock it was always temperature that limited the process.

[bogusmill]

On this review it was water cooled but it comes with water cooling hardware. Temp is limiting mine to but it's the northbridge not the CPU. If NB gets over 55C a core errors out in Prime 95. Will be addressing this soon, maybe then I can pump up the volume a little.

[Guest CNC Apps Guy 1]

Can you go with LOX or LN. That ought to cool 'er right down...

[bogusmill]

quote:

---

Can you go with LOX or LN

---

Actually yes we can. In research tests for high altitude we use LN and there are lots of LN tanks scattered about NIAR. Unfortunately we're forced by state contract to buy Dell or HP. That board on LN has run at 6.4+Gb, catch me if you can.

[Rolling Paper]

9 minutes 29 seconds

 

Hp xw9300 Workstation

AMD Opteron 250 (64 bit) @ 2.39 GHz

Windows XP Pro X64

RAM: 2048 MB

Quadro FX 3500

[Guest]

Trying to get stable at 3.375Ghz

 

Juuuuust about there.

 

This new motherboard is sweet.

 

I hit 3.2 Ghz with just about stock voltages.

 

Trying now at 375 FSB x 9, have my RAM current at 750 MHZ for a 1:1 ratio, after I stabilize the CPU, I'll tweaking the RAM up.

[Guest]

As Bogusmill reuqested

 

 

3:45 seconds

[bogusmill]

Way cool John. Getting to the point where tenth's of a second seperate the best times.

 

My 8GB RAM kit will be here Tue. for my Vista x64 partition. I'm doing some cooling mods & updating the BIOS too. I'm also adding a 4GB RAM drive to my XP x86 boot partition to make use of the excess RAM. Maybe I can close some of that 15 second lead you have.

 

Might have found another secret weapon. I'll tell you if it works, for now I'll only tell you it involves a Penguin.