Krzysztof_11 :
Dear All,
my wife is doing some interior design projects in 3ds max (with V-Ray). We decided to build her a new PC so her work will go more smoothly. I was thinking about:
- Intel Core i7-7700K 4.2 GHz
- Gigabyte GeForce® GTX 1070 G1 Gaming, 8GB
- Motherboard MSI Z270-A PRO
- Corsair Vengeance LPX 2x8GB 2400MHz DDR4 Intel XMP 2.0 (I will double it up asap)
- SSD Intel 600p series M.2 PCIe 256GB
- HDD Seagate BarraCuda 2TB 64MB 7200
What do you think about this configuration? Shall I look for some better CPU and cut the GPU? Or did 3ds max will still use the GPU so it still has to be good?
Also is it possible to use M.2 SSD on Windows 7 or shall I stick to some SATA SSD like Samsung EVO 850?
I will be very grateful for any answers
Krzysztof_11,
The i7-7700K is a very good choice for 3D modeling. The CPU's single-thread performance is the important factor and the i7-7700K has the 2nd highest Passmark Single Thread Mark of
2600. For comparison, the E5-1660 v2 I use for 3D modeling has the second highest Xeon E5 Single Thread Mark at
2077.
However, for VRay rendering, core count can be more important and the i7-7700K is a 4-core /8-thread processor. It's possible to run the VRay RT engine in GPU mode, and that is very fast, but in my view, GPU rendering quality is inferior to CPU rendering. This is seen in GPU renderings as rainbowed color gradients, raw edges, lack of subtlety in shadow transitions, "lifeless" reflections and water, and to me a slight general sense of artificiality. This is especially important in interiors as the views are close to the objects and there are multiple lighting sources.
CPU rendering is however, very processor intensive. Because of this, since 2010 or so, I've had separate systems for modeling and CPU rendering. a few days I tried a CPU VRay test rendering (3180 X2160) on a 152MB Sketchup model on the modeling system with a Xeon E5-1660 V2 6-core@3.7 /4.0GHz /32GB of RAM/ Quadro K4200 4GB and it failed twice. The rendering system: 2X Xeon E5-2690 8-core#2.9/3.8GHz /64GB/Quadro K2200 + Tesla M2090 ran all 32threads at 100% and finished in 10min 30secs. the key was that the rendering processes was using 37GB of RAM- the 32GB E5-1660 V2 system didn't have enough.
This is not to say that the renderings being run on the proposed system will need a 32-thread system only that they will definitely have to be GPU-based and 16GB is the absolute starter RAM set- if possible, start out with 32GB of RAM. If carefully set up, GPU renderings can be acceptable quality.
If the rendering is to be GPU-based, then the GTX 1070 is an extremely fast GPU and very good value. I use Quadros for the viewport support and higher anti-aliasing, multiple lighting sources, 30-bit color, and etc. For a content creation workstation, you might consider the upcoming- next month- Quadro P2000 5GB which runs at about GTX 780 speeds. That will cost something under $500. Yes, it is quite a bit slower than a GTX, but the difference in rendering time of single images may be very little. I tried the identical Sketchup/VRay RT GPU rendering (30MB model on the E5-1660 v2 /Quadro K4200 and it was perhaps 1:30 faster than the CPU version. GPU rendering really excels in video /animations where there are thousands of frames. If there is to be video editing /processing then by all means the GTX 1070 is a great one.
An M.2 drive is in general a good idea, and I use one- Samsung SM951 256GB in the E5-2690 rendering system for the OS/programs and the projects run off a Samsung 850 Evo 250GB. It would probably be faster to contain the active projects on the M.2. On the E5-1660 v2 system, the drive is an Intel 730 480GB SATA SSD with a partition that contains all files. In practice I don't really notice any speed difference between the two SSD's. I've casually timed certain processes with a stopwatch and the system starting on M.2 is about 8 sec.faster, opening a large file is about 4 secs. faster, saving that file is perhaps 8-10 secs faster, and so on. This has to be that the CPU is throttling the M.2 speed. An all SSD system would be different.
The brilliance of M.2 is in large file transfers, but as I use WD Black and Seagate ES.3 drives , the M.2 transfer speed is wasted as the mech.l hard drives can't write at 1/10th M.2 speeds - a bit better with large media files. The M.2 looks great in benchmarks- the E5-2690 has scored
13460 on Passmark whereas the Intel 730 is
4794, but I've yet to really "feel" M.2 is any faster and you might consider instead having a Samsung 850 Evo 500GB and create a 150GB OS/ Programs partition and 350GB for all files. All the files I've created since 1993- about 30,000 images, 1,400 CAD drawings, 400 3D CAD models, 22,000 documents fit in 72GB.
For stability, don't overclock.
Cheers,
BambiBoom
CAD / 3D Modeling / Graphic Design:
HP z420 (2015) (Rev 3) > Xeon E5-1660 v2 (6-core @ 3.7 / 4.0GHz) / 32GB DDR3 -1866 ECC RAM / Quadro K4200 (4GB) / Samsung SM951 M.2 256GB AHCI + Intel 730 480GB (9SSDSC2BP480G4R5) + Western Digital Black WD1003FZEX 1TB> M-Audio 192 sound card + Logitech z2300 2.1 speakers > 600W PSU> > Windows 7 Professional 64-bit >> 2X Dell Ultrasharp U2715H (2560 X 1440)
[ Passmark Rating = 5581 > CPU= 14046 / 2D= 838 / 3D= 4694 / Mem= 2777 / Disk= 11559] [6.12.16]
Analysis / Simulation / Rendering:
HP z620 (2012) (Rev 3) 2X Xeon E5-2690 (8-core @ 2.9 / 3.8GHz) / 64GB DDR3-1600 ECC reg) / Quadro K2200 (4GB) + Tesla M2090 (6GB) / HP Z Turbo Drive (256GB) + Samsung 850 Evo 250GB + Seagate Constellation ES.3 (1TB) / Creative Sound Blaster X-Fi Titanium PCIe sound card + Logitech z313 2.1 speakers / 800W / Windows 7 Professional 64-bit > > HP 2711x (27" 1980 X 1080)
[ Passmark System Rating= 5675 / CPU= 22625 / 2D= 815 / 3D = 3580 / Mem = 2522 / Disk = 12640 ] 9.25.16
[ Cinebench R15: OpenGL= 119.23 fps / CPU = 2209 cb / Single core 130 cb / MP Ratio 16.84x] 10.31.16