thanks a lot. that helped me get some clear thoughts again. after reading so much about the topics i didn't spend any time with in the past 10 years, i was more confused than informed.
Hello everyone, I am new to this forum and I would appreciate some help choosing a laptop suitable for 3d modeling work. https://www.neptun.mk/categories/GAMING_LAPTOPI/ACER-NITRO-5-AN517-41-R8MN-17-3--R7-5800H-16GB-512GB-RTX-3060-6GB Specifically I need an opinion for this one. From what i saw available where i live, this one looked like the best deal for my budget. I would really appreciate your help. ACER NITRO 5 AN517-41-R8MN 17.3"
R7-5800H/16GB/512GB/RTX 3060 6GB
Hi, i've got 3070 which i am using for some time. I dont really like the amount of vram on this card, should i swap for 3080ti with additional payment or swap for 2x 3060? Mostly blender archviz
While multi gpu setups can be useful for rendering, keep in mind that the viewport itself can only use one card at a time. Which aspect of your workflow is suffering from the lack of vram?
While multi gpu setups can be useful for rendering, keep in mind that the viewport itself can only use one card at a time. Which aspect of your workflow is suffering from the lack of vram?
Rendering mostly as scenes are quite heavy. But smooth vp is important to me. I have been using dual 1080ti for some time and i am afraid i will go back to this kind of performance despite vram boost
The main thing you want to pay attention to is color accuracy / gamut coverage. Both of those have a 90% or better DCI-P3 coverage. Basically DCI-P3 is a new color accuracy standard that fully encompasses older standards like sRGB and Adobe RGB, while being suitable for measuring how accurate the screen will be when dealing with HDR content.
The other big thing to worry about is light bleed, which is primarily a quality control issue. Basically if you set the screen color to be black and it isn't black (some parts are brighter then others) that means you have light bleed, and it can/will mess with your color perception. Acer produces cheap 4k 120hz screens with good color accuracy for instance that you could get for half the price of the ones you mentioned, but light bleed is extremely common on them.
Note that if you want to run 4K at high refresh rates (120hz or higher) you'll either need a DSC capable video card (basically a form of video compression), or run it at a lower color accuracy (8-bit instead of 10-bit). So if you're still using a Geforce 1000 series GPU or older you'll need to upgrade that to get the full benefits of the monitor.
hey guys, i'm currently running a GTX 1060 6GB (and sometimes feeling the chug hard) and i'm looking to buy an RTX 3060 or 3060ti, but i don't know if the performance difference in Substance Painter will be big enough to warrant buying it now during the price hike. I'm not concerned with baking speed or rendering, all i'm looking for is faster response times when working with masks, mats, and viewport update speed when texturing. What do guys think? Should i just bite the bullet or should i wait it out with my 1060?
hey guys, i'm currently running a GTX 1060 6GB (and sometimes feeling the chug hard) and i'm looking to buy an RTX 3060 or 3060ti, but i don't know if the performance difference in Substance Painter will be big enough to warrant buying it now during the price hike. I'm not concerned with baking speed or rendering, all i'm looking for is faster response times when working with masks, mats, and viewport update speed when texturing. What do guys think? Should i just bite the bullet or should i wait it out with my 1060?
The prices have gotten so high on gpu prices in germany, I could litteraly buy a laptop with a rtx 3070 in it for the same price of a desktop RTX 3070 (gpu):P
hey guys, i'm currently running a GTX 1060 6GB (and sometimes feeling the chug hard) and i'm looking to buy an RTX 3060 or 3060ti, but i don't know if the performance difference in Substance Painter will be big enough to warrant buying it now during the price hike. I'm not concerned with baking speed or rendering, all i'm looking for is faster response times when working with masks, mats, and viewport update speed when texturing. What do guys think? Should i just bite the bullet or should i wait it out with my 1060?
The prices have gotten so high on gpu prices in germany, I could litteraly buy a laptop with a rtx 3070 in it for the same price of a desktop RTX 3070 (gpu):P
I know, it's crazy. It's not quite as bad in Russia, i can still snag a 3060 that is only around 3 times more expensive than it should be. What a deal!
Well I've not touched a 3060/ti, but we've got 3080s and 1080tis in the studio, I've also got a 1060 6GB at home, It's not really a great comparison, but judging by benches of a 3060ti it's probably near around a 1080ti performance. I would estimate it's a pretty good upgrade for slow downs in painter, the 1080ti usually didn't get many gripes about performance in painter.
Though "Waiting it out"... waiting what out? Maybe the market has decided that gpus are sustainable at these very high retail prices and you can no longer get a decent gpu in £300 range unless it's 2 gens old.
I'm kinda desperate in remaining hopeful that the market will go back at least somewhat (it will, right?? RIGHT?) at some point in the future. In the end, though, I've decided against getting the 3060. Just as a point of principle against paying three times the prices for something. I'm sure that will show them.
I wish I had gotten a 3080 over the RX 6800 I have (while I love it, Working in SP with an AMD card, god forbid if that worked....well unless you have the new SP with subscription.. which I Refuse to do ).
Hey. I’m upgrading my old i7 6800k 4 core 8 thread pc. I’m looking to get either a 3900x or 3950x. (Not sure I want to go for the newer gen ryzen as I want to keep as much of my build and just upgrade what I need).
My current build works perfect for all the usual stuff (modelling, zbrush/texturing) it just doesn’t cut it when I want to render marginally intense scenes in Arnold with Udims/xgen/sss etc.
I'm thinking of getting the Ryzen 3900x, 64gb (3200), Noctua D15 fan, Asus Rog Strix x570-F MOBO, the rest I already have. Is the 3950x worth the extra 4 cores and thus the extra £250/is there a noticeable difference in speed with the extra cores for rendering? Suppose anythings better than my current build's modest 4 cores haha
Unless you found a really good deal on then, there's basically no reason to choose the Ryzen 3000 series over the 5000. Also keep in mind that the Ryzen 6000 series (or whatever they end up called the ones with 3D V-Cache) are rumored to launch in about 2-3 months, and those should be the final CPUs for the AM4 platform before AM5 takes over.
My PC is about 9 years old and it's been long due for an upgrade. I already bought an EVGA RTX 2060 KO Ultra last year, so I'm just upgrading my CPU, mobo, and RAM now. I'm mainly 3D modeling, texturing, and rendering. I bought the mobo already on Amazon and I'm gonna pick up the CPU + RAM soon from Micro Center. Just wondering how I did in my selection:
G.Skill Ripjaws V 2x16 GB DDR4-3600 CL18-22-22-42 : $117.99
Total: $466.98 + tax
I wanna keep the total under $500 and the CPU is the centerpiece due to Black Friday sales. My i5-3570k only has four cores/threads, so the 12 threads on the 5600G should be a nice boost for occasional tasks like using combined CPU+GPU mode in UVPackmaster and CPU-only OSL bevel shader baking in Blender. The fact that it comes with a free cooler is such great value. The mobo seems to be highly rated in previous and current iterations with great VRMs according to this thread, so I should be able to give the CPU a decent overclock without worries. I currently have 16 GB RAM and find myself running out when having different DCCs and a ton of Firefox tabs open, so I figured I'd go 32 GB this time. I'm used to that particular brand of RAM and the price seems about average.
Happy holidays and I hope you guys find some great deals too.
The 5600G is quite a bit slower then the 5600X, but at that price you did ok. For the RAM, make sure after installing to go into the system bios and enable XMP so it runs at full speed:
@PolyHertz Yeah I'm definitely a value-minded buyer. The 5600X alone is $279.99 at Micro Center when not bundling a qualifying mobo, and that's apparently the lowest price it's been. It's a 40% price increase over the 5600G though. I'm looking at benchmarks here and it seems the 5600X is between 11 to 15% faster in CPU-only rendering. I agree that's a pretty good difference, but idk if I can justify the 40% price increase for it.
Thank you for the suggestions. I'll have to look into whether Precision Boost Overdrive works well or if I'll have to do a manual OC. Hopefully I can just migrate Windows 10 by plugging my drives into the new mobo too lol.
Edit: The 5600X is now $250 at Micro Center and I could do an exchange. I'm mainly eyeing the L3 cache and I wonder how much of a difference 32 vs 16 MB would be. I don't have any PCIe 4.0 devices currently.
Hi everyone. Thanks for the thread and all the info you provide specially @PolyHertz I've been lurking and I'm learning a ton but there are a lot of things still confusing to me. For example, if Substance Painter uses GPU memory, does that mean that an RTX 3060, 12GB VRAM would be a better bang for buck option for asset creation than 3060 TI - 3070 TI (8GB), or 3080 (10GB)?
I'm a character artist who wants to specialize in current gen, hyper realistic, real time game assets and I'm planning to build a new rig at least 6 months from now and hopefully prices would be better by then. I mostly use Zbrush, Marvelous Designer, Substance Painter, 3ds Max, Unreal Engine (for lookdev and presentation characters, no plans to do levels) I do not intend to play games on this machine. I plan to use this machine for at least 6 years before considering upgrading/replacing
I'm looking at: Ryzen 5900x MSI MAG 550 Tomahawk Noctua NH-D15 GSKILL TridentZ 2x16GB DDR4 3600 C16D Samsung 970 Evo Plus m.2 NVME 1 TB WD Caviar Blue 4TB EVGA Supernova 750 GA Gold Lianli Lancool II Performance X RTX 30?? - This one I would love to hear your thoughts for the best bang for buck for what I do. If I go for a 5900x processor I can probably go between RTX 3060-3070. But if you guys think I should go for at least 3070 TI or 3080 would it make sense if I save on the CPU and go for a Ryzen 3600 or 5600x?
I'm open to suggestions on other parts as well for more optimization if you guys think this build I'm planning is too much for what I intend to do with it. My current machine is an almost 6 year old i7-6700, gtx 1070, 16 gig ram. Not sure if this helps.
Right now Epic recommends for UE5 a 12 core CPU (Ryzen 5900X), 64GB of RAM, and a Geforce 1080 or better GPU (at least for running the "Valley of the Ancients" demo). So you probably should get a PC with at least those specs atp if you want something that will last.
For GPUs, while the 3060 does have the most vram compared to the 3060Ti/3070/3080, it also has a slower memory bus which negates a lot of the benefits of having more vram. That said it's a decent enough card, and should be OK for doing UE5 and Substance work.
Also, you said you're not building for at least 6 months, and by then the final AM4 CPUs should be released (5900X with V-Cache), while late 2022 should see the release of a whole new AMD chipset (AM5) and the Zen4 series of CPUs. There are rumors that Nvidia will be launching the Geforce 4000 series at the end of 2022 as well (the 3000 series is over a year old atp after all). So be sure to check on whats available or soon to be available before pulling the trigger on your purchase.
Is there any practical advantage for our work that the AMD 5950 might have over the 5900 other than the number of cores? Any performance or functional difference when paired with 128 GB of RAM?
Also one thing I have done on my current machine and want to keep doing on the new one is to have a clone for the Windows boot drive. Boot drive in my case is entirely separate from data/work and only contains the OS and applications. Also there's only a single Windows install on the machine, nothing else.
I want to be able to clone my boot drives back and forth, hook one up and disconnect the other and have the option to stick them into an external case if needed for troubleshooting. All the while it should make zero difference for Windows where it is being booted from so long as I pick the right boot device in the boot manager/BIOS. Question is also if I should be going with SATA SSDs (cheaper and in my experience performance-wise sufficient for the task) or NVMe's.
For this purpose is there anything to consider in regards to AMD/mainboards/UEFI to be used with Windows 10 (not 11 - for now)?
There are no functional differences between the 5900X and 5950X, just more cores and a bit more L2 cache (8MB on the 5950X instead of 6MB). The additional 8 threads can of course help speed up compiling / light baking / CPU -based rendering operations, but it's not going to be a world changing difference.
As for NVMe vs SATA, the main benefits atm are realtime video scrubbing speeds (if you're editing high resolution videos), or if you're regularly writing/moving large files. I've also heard it helps when used as a cache for simulations. Not much benefit beyond those things for now, although if Microsofts 'Direct Storage' API starts getting used then game performance may also begin being affected.
My question re:SSDs is more about any restrictions to look out for in terms of secure boot/UEFI stuff that might be in these new fancy mainboards. Basically is there anything that might restrict me from making a straight clone of the OS SSD and plugging that one in - internally or externally when using NVMe as opposed to SATA?
Asking because at least on the Mac that is a real sh*tshow these days. Sometimes more practical to buy the older, dumber tech. /sigh
Thanks, yeah just checked the RAM and it's latency 16. Phew ;) I also included the 5950X in the end because it still fit the budget, so might as well get extra cores.
The computer is currently being assembled by a shop, I hope it delivers. Should have it end of the week.
Quick question - NVidia studio driver or game-ready driver - which to pick? Studio seems like the obvious choice but do we know of any issues with Unreal or other engines/editors? (I don't play any actual games on PC) .
Now after a few weeks in full production mode with the new system I have to admit I don't perceive the Ryzen 5950X as any faster than my 8-year old Intel I upgraded from for single core heavy tasks (so, basically everything that isn't rendering, baking, cooking things in Unreal or zipping up archives). On paper there should be at least a 40% increase but I'm not feeling it.
Perhaps if anybody was still willing to pay me for sculpting I'd see the benefits by quadrupling cores but it seems grooming is where it's at for me and honestly there's no difference in interaction speed that I could perceive without a stop watch. I knew my old i7-4790K was one of the top single core scorers for a looooong time but... well, I would have expected more from the Ryzen and the all-new computer guts surrounding it.
Instead I got a bunch of new things to worry about, like managing temperatures or rather fan speeds on that permanently turbo-boosting CPU that by default made the cooling system spin up and down all the time even when just starting up programs. Then there's tools like ClockTuner to opti....complicate things further.
That must be what they call progress, wonderful. 🤪
Oh thanks but its not a Maya-specific comment. Everything I tend to do all day long - modelling, uv-mapping or using a paint software or something like Toolbag. It's not really reacting any faster. There's no noticeable single core speed boost that tells you: new machine, everything is flying now.
I took an endlessly long time to configure the new computer and move everything over simply because the old one wasn't feeling inferior except for the RAM and I preferred to work instead of fiddle with Windows configuration.
Unfortunately anyone whose used a 4+ core CPU from Sandy Bridge (2500K and 2600K) on up, with an SSD, has basically already experienced the final "Oh shit! Everything's so much faster then on my last PC!" moment (at least for the foreseeable future). A 40% boost in per-thread performance is nothing, it's like a game going from 30fps to 42fps, most people wont even notice. The main reason these days to buy a new CPU/chipset is simply for more cores and more ram. If the software you use isn't designed to take advantage of those extra resources then there's really nothing you can do other then experiment with different software to see if anything else gives you the features + performance you're looking for.
That doesn't sound good at all. I have an AMD 3600. Should I expect little in the way of performance of a 5950x upgrade? I tend to do most of my work in Photoshop and working on 45MP camera raw is sluggish.
I also wonder if the 5800X3D is the very last iteration of the AM4 platform based CPU.
@phaedarus For Photoshop performance in general; In the newest releases of Photoshop you can speed up performance in some instances by going to Preferences > Performance and enable "Multithreaded compositing" as well as "Use Graphics Processor". You should also make sure you have and are allocating enough RAM to Photoshop there, and the correct cache levels / tile size.
As for using a newer hardware; different features (such as the various filters) can use a varying number of threads. So some features will benefit from a newer CPU with more threads and others wont. Using lower latency ram may also help as seen here. (basic latency formula; 2000 * CL / frequency , so DDR4 2600 CL16 = 13.3ns of latency , while DDR4 3600 CL16 == 8.8ns)
Agreed, seems like the last time I felt a proper performance boost was when going from Core2 to i7. But with GPU upgrades I certainly do notice something like a 40% boost. Oh well, I wasn't comfortable with running an 8-year old as my daily driver anymore so I'm not too fussed. Not exactly a necessary upgrade though.
What it definitely will affect is my upgrade path for my other desktop - a Mac Pro Trashcan that is slightly faster than my old 4790k. I think that will stay with me much longer now than anticipated. I expected it to be crushed by the AMD but as it turns out... *shrug*
@phaedarus I think you might as well look at other software to see if that makes a difference. Affinity Photo should have the tools you need and in my experience is generally faster than PS. Not too familiar with the very latest PS versions though. Also perhaps the file format you are using is just slow to handle and things might work better if converted to application native format?
Yeah they're checked. I have the default 70GB of memory allocated. I believe my RAM is 3200mhz CL18 DDR4 using the MSI MOBO profile to increase that to 3600mhz. I just don't understand why Photoshop causes so much lag on my system that even the application switcher slows down (alt-tab). I have a separate 1TB M.2 NVME drive just for cache. No other app behaves this badly.
Do you think it's reasonable to switch from a 3900 to 5950x ? I'll do essentially 2D illustrations, 3D modeling, texturing (SP), baking (SP, toolbag) and Unity/Unreal integration ( with lighting). I can get one for 460€. With AM5 which start in fall, i'm still hesitating to make the step.
In same time, the new architecture would be far more expensive (CPU :500-600€ for 12C/24T- mobo : 130-150€ - 32Go DD5: 250-300€).
For any texture or light baking that uses the CPU it should provide a very noticeable performance improvement.
Game engines like Unreal and Unity will have more consistent framerates due to the better cache design of Zen 3 (Ryzen 5000 series) over Zen 2 (Ryzen 3000 series).
Working in 2D, or standard 3D modeling, are more single threaded tasks that wont really benefit much from it.
Hey guys super random question, I am trying to create a new pc for CG/3D work. I have been reading lots of articles on which processors are the "best" for CG and so these articles have made two categories on which process you should get "active work" and "Render work". I am mainly going to be using this pc for active work for modeling sculpting etc but I also will be rendering on this pc. They are saying to get the intel 12900k for the active work and a threadripper for rendering work. Since the begging I have been looking for a threadripper and was wondering if having the threadripper will lower software performance since it has many cores but they are slow and most softwares use single core. Please let me know on what you guys think I should do I am very lost.
The PC I plan to build, for modelling, texturing, sculpting in Maya, Substance Painter/Designer, Zbrush, Unity and Unreal 4 & 5.
Intel® Core™ i9-12900KF Processor
Asus Z690-Plus TUF Gaming bundkort
2 X Kingston Fury Beast 32GB DDR5-5200 RAM (64gb)
Western Digital Black 1TB SN750SE PCle 4.0 NVMe™ SSD
Kingston NV1 2TB M.2 NVMe SSD
Asus GeForce® RTX 3080 Ti 12GB TUF
Any tips or obvious drawbacks?
Would it be worth sacrificing some components, like less or slower RAM, removing a SSD etc to upgrade from the RTX 3080 TI 12GB to the RTX 3090 TUF 24GB?
It should be fine for standard modeling and animation tasks, just don't expect to be doing any rendering with it (maybe look into offloading any renders to a cloud service). Also make sure to install an SSD in it if it doesn't already have one.
Replies
https://www.neptun.mk/categories/GAMING_LAPTOPI/ACER-NITRO-5-AN517-41-R8MN-17-3--R7-5800H-16GB-512GB-RTX-3060-6GB
Specifically I need an opinion for this one. From what i saw available where i live, this one looked like the best deal for my budget.
I would really appreciate your help.
ACER NITRO 5 AN517-41-R8MN 17.3" R7-5800H/16GB/512GB/RTX 3060 6GB
Display:17.3" (43.9cm)
(1920x1080) FHD IPS 144 Hz
CPU:AMD Ryzen R7-5800H 3.20 GHz Octa-core (8 Core™)
GPU:NVIDIA GeForce RTX 3060 6GB
RAM:16GB DDR4
Memory :512GB SSD
Mostly blender archviz
The main thing you want to pay attention to is color accuracy / gamut coverage. Both of those have a 90% or better DCI-P3 coverage. Basically DCI-P3 is a new color accuracy standard that fully encompasses older standards like sRGB and Adobe RGB, while being suitable for measuring how accurate the screen will be when dealing with HDR content.
The other big thing to worry about is light bleed, which is primarily a quality control issue. Basically if you set the screen color to be black and it isn't black (some parts are brighter then others) that means you have light bleed, and it can/will mess with your color perception. Acer produces cheap 4k 120hz screens with good color accuracy for instance that you could get for half the price of the ones you mentioned, but light bleed is extremely common on them.
Note that if you want to run 4K at high refresh rates (120hz or higher) you'll either need a DSC capable video card (basically a form of video compression), or run it at a lower color accuracy (8-bit instead of 10-bit). So if you're still using a Geforce 1000 series GPU or older you'll need to upgrade that to get the full benefits of the monitor.
i'm currently running a GTX 1060 6GB (and sometimes feeling the chug hard) and i'm looking to buy an RTX 3060 or 3060ti, but i don't know if the performance difference in Substance Painter will be big enough to warrant buying it now during the price hike. I'm not concerned with baking speed or rendering, all i'm looking for is faster response times when working with masks, mats, and viewport update speed when texturing. What do guys think? Should i just bite the bullet or should i wait it out with my 1060?
The prices have gotten so high on gpu prices in germany, I could litteraly buy a laptop with a rtx 3070 in it for the same price of a desktop RTX 3070 (gpu):P
Though "Waiting it out"... waiting what out? Maybe the market has decided that gpus are sustainable at these very high retail prices and you can no longer get a decent gpu in £300 range unless it's 2 gens old.
My current build works perfect for all the usual stuff (modelling, zbrush/texturing) it just doesn’t cut it when I want to render marginally intense scenes in Arnold with Udims/xgen/sss etc.
I'm thinking of getting the Ryzen 3900x, 64gb (3200), Noctua D15 fan, Asus Rog Strix x570-F MOBO, the rest I already have. Is the 3950x worth the extra 4 cores and thus the extra £250/is there a noticeable difference in speed with the extra cores for rendering? Suppose anythings better than my current build's modest 4 cores haha
For the RAM, make sure after installing to go into the system bios and enable XMP so it runs at full speed:
https://www.youtube.com/watch?v=orENI_3nHos
Thanks for the thread and all the info you provide specially @PolyHertz
I've been lurking and I'm learning a ton but there are a lot of things still confusing to me. For example, if Substance Painter uses GPU memory, does that mean that an RTX 3060, 12GB VRAM would be a better bang for buck option for asset creation than 3060 TI - 3070 TI (8GB), or 3080 (10GB)?
I'm a character artist who wants to specialize in current gen, hyper realistic, real time game assets and I'm planning to build a new rig at least 6 months from now and hopefully prices would be better by then. I mostly use Zbrush, Marvelous Designer, Substance Painter, 3ds Max, Unreal Engine (for lookdev and presentation characters, no plans to do levels)
I do not intend to play games on this machine. I plan to use this machine for at least 6 years before considering upgrading/replacing
I'm looking at:
Ryzen 5900x
MSI MAG 550 Tomahawk
Noctua NH-D15
GSKILL TridentZ 2x16GB DDR4 3600 C16D
Samsung 970 Evo Plus m.2 NVME 1 TB
WD Caviar Blue 4TB
EVGA Supernova 750 GA Gold
Lianli Lancool II Performance X
RTX 30?? - This one I would love to hear your thoughts for the best bang for buck for what I do.
If I go for a 5900x processor I can probably go between RTX 3060-3070.
But if you guys think I should go for at least 3070 TI or 3080 would it make sense if I save on the CPU and go for a Ryzen 3600 or 5600x?
I'm open to suggestions on other parts as well for more optimization if you guys think this build I'm planning is too much for what I intend to do with it. My current machine is an almost 6 year old i7-6700, gtx 1070, 16 gig ram. Not sure if this helps.
Thanks so much.
Right now Epic recommends for UE5 a 12 core CPU (Ryzen 5900X), 64GB of RAM, and a Geforce 1080 or better GPU (at least for running the "Valley of the Ancients" demo). So you probably should get a PC with at least those specs atp if you want something that will last.
For GPUs, while the 3060 does have the most vram compared to the 3060Ti/3070/3080, it also has a slower memory bus which negates a lot of the benefits of having more vram. That said it's a decent enough card, and should be OK for doing UE5 and Substance work.
Also, you said you're not building for at least 6 months, and by then the final AM4 CPUs should be released (5900X with V-Cache), while late 2022 should see the release of a whole new AMD chipset (AM5) and the Zen4 series of CPUs. There are rumors that Nvidia will be launching the Geforce 4000 series at the end of 2022 as well (the 3000 series is over a year old atp after all). So be sure to check on whats available or soon to be available before pulling the trigger on your purchase.
What effect, if any, does the AMD 5900x with 3D cache have on DCC software?
My original plan was to upgrade my 3600 to 5950x but I would like to get your take on this latest refresh before I do so.
New computer time! First AMD in two decades.
Is there any practical advantage for our work that the AMD 5950 might have over the 5900 other than the number of cores? Any performance or functional difference when paired with 128 GB of RAM?
Also one thing I have done on my current machine and want to keep doing on the new one is to have a clone for the Windows boot drive. Boot drive in my case is entirely separate from data/work and only contains the OS and applications. Also there's only a single Windows install on the machine, nothing else.
I want to be able to clone my boot drives back and forth, hook one up and disconnect the other and have the option to stick them into an external case if needed for troubleshooting. All the while it should make zero difference for Windows where it is being booted from so long as I pick the right boot device in the boot manager/BIOS. Question is also if I should be going with SATA SSDs (cheaper and in my experience performance-wise sufficient for the task) or NVMe's.
For this purpose is there anything to consider in regards to AMD/mainboards/UEFI to be used with Windows 10 (not 11 - for now)?
Thanks!
There are no functional differences between the 5900X and 5950X, just more cores and a bit more L2 cache (8MB on the 5950X instead of 6MB). The additional 8 threads can of course help speed up compiling / light baking / CPU -based rendering operations, but it's not going to be a world changing difference.
As for NVMe vs SATA, the main benefits atm are realtime video scrubbing speeds (if you're editing high resolution videos), or if you're regularly writing/moving large files. I've also heard it helps when used as a cache for simulations. Not much benefit beyond those things for now, although if Microsofts 'Direct Storage' API starts getting used then game performance may also begin being affected.
Thanks, I'll be picking the 5900 then!
My question re:SSDs is more about any restrictions to look out for in terms of secure boot/UEFI stuff that might be in these new fancy mainboards. Basically is there anything that might restrict me from making a straight clone of the OS SSD and plugging that one in - internally or externally when using NVMe as opposed to SATA?
Asking because at least on the Mac that is a real sh*tshow these days. Sometimes more practical to buy the older, dumber tech. /sigh
Ok, my config has been put together now:
Case: Fractal Design R7
Case fans: beQuiet 140mm Silent Wings 3 (3x)
CPU: AMD Ryzen 9 5950X
CPU Cooler: Noctua NH-D15S
Mainboard: Gigabyte X570S Aorus Master
GPU: Asus ROG Strix 2080 Ti (moved from current machine, probably gonna skip the RTX 3xxx-series entirely)
RAM: 128 GB DDR4-3200 Corsair Vengeance RGB Pro
Power Supply: beQuiet Straight Power 11 Platinum 750W
SSD: Samsung 980 Pro PCIE 4.0 - work
SSD: Samsung 870 EVO (2x) - OS + OS-clone
+ a few still pretty new SSDs and one HDD moved from current machine for archives and backup
Late to reply, but just wanted to say that looks like a great build!
Just make sure the RAM has a cas latency of 16 or lower to avoid potential frame spikes / stutter (if you haven't already bought it).
Thanks, yeah just checked the RAM and it's latency 16. Phew ;) I also included the 5950X in the end because it still fit the budget, so might as well get extra cores.
The computer is currently being assembled by a shop, I hope it delivers. Should have it end of the week.
Quick question - NVidia studio driver or game-ready driver - which to pick? Studio seems like the obvious choice but do we know of any issues with Unreal or other engines/editors? (I don't play any actual games on PC) .
@thomasp Studio drivers are geared for helping create stable environments for content creation.
I just use game-ready as I do play and work on the same rig.
Just updated the ram "What to know" section, adding some extra info about latency and overclocking for those that might be interested.
Whatever happened to the V Cache for the 5950x?
Unfortunately AMD scrapped plans for V-Cache on all Zen3 based CPUs other then the 5800X (which they're calling the 5800X3D).
Now after a few weeks in full production mode with the new system I have to admit I don't perceive the Ryzen 5950X as any faster than my 8-year old Intel I upgraded from for single core heavy tasks (so, basically everything that isn't rendering, baking, cooking things in Unreal or zipping up archives). On paper there should be at least a 40% increase but I'm not feeling it.
Perhaps if anybody was still willing to pay me for sculpting I'd see the benefits by quadrupling cores but it seems grooming is where it's at for me and honestly there's no difference in interaction speed that I could perceive without a stop watch. I knew my old i7-4790K was one of the top single core scorers for a looooong time but... well, I would have expected more from the Ryzen and the all-new computer guts surrounding it.
Instead I got a bunch of new things to worry about, like managing temperatures or rather fan speeds on that permanently turbo-boosting CPU that by default made the cooling system spin up and down all the time even when just starting up programs. Then there's tools like ClockTuner to
opti....complicate things further.That must be what they call progress, wonderful. 🤪
@thomasp If you're using xgen for hair you might need to enable multi-threading in preferences take advantage of extra cores, at least according to this: https://knowledge.autodesk.com/support/maya/learn-explore/caas/CloudHelp/cloudhelp/2018/ENU/Maya-Customizing/files/GUID-60D167E6-7A25-4B4F-8C7C-1117181ED79E-htm.html
(I don't have Maya installed atm to check)
Oh thanks but its not a Maya-specific comment. Everything I tend to do all day long - modelling, uv-mapping or using a paint software or something like Toolbag. It's not really reacting any faster. There's no noticeable single core speed boost that tells you: new machine, everything is flying now.
I took an endlessly long time to configure the new computer and move everything over simply because the old one wasn't feeling inferior except for the RAM and I preferred to work instead of fiddle with Windows configuration.
@thomasp Ah, yea I get what you're saying.
Unfortunately anyone whose used a 4+ core CPU from Sandy Bridge (2500K and 2600K) on up, with an SSD, has basically already experienced the final "Oh shit! Everything's so much faster then on my last PC!" moment (at least for the foreseeable future). A 40% boost in per-thread performance is nothing, it's like a game going from 30fps to 42fps, most people wont even notice. The main reason these days to buy a new CPU/chipset is simply for more cores and more ram. If the software you use isn't designed to take advantage of those extra resources then there's really nothing you can do other then experiment with different software to see if anything else gives you the features + performance you're looking for.
That doesn't sound good at all. I have an AMD 3600. Should I expect little in the way of performance of a 5950x upgrade? I tend to do most of my work in Photoshop and working on 45MP camera raw is sluggish.
I also wonder if the 5800X3D is the very last iteration of the AM4 platform based CPU.
@phaedarus For Photoshop performance in general; In the newest releases of Photoshop you can speed up performance in some instances by going to Preferences > Performance and enable "Multithreaded compositing" as well as "Use Graphics Processor". You should also make sure you have and are allocating enough RAM to Photoshop there, and the correct cache levels / tile size.
As for using a newer hardware; different features (such as the various filters) can use a varying number of threads. So some features will benefit from a newer CPU with more threads and others wont. Using lower latency ram may also help as seen here. (basic latency formula; 2000 * CL / frequency , so DDR4 2600 CL16 = 13.3ns of latency , while DDR4 3600 CL16 == 8.8ns)
Agreed, seems like the last time I felt a proper performance boost was when going from Core2 to i7. But with GPU upgrades I certainly do notice something like a 40% boost. Oh well, I wasn't comfortable with running an 8-year old as my daily driver anymore so I'm not too fussed. Not exactly a necessary upgrade though.
What it definitely will affect is my upgrade path for my other desktop - a Mac Pro Trashcan that is slightly faster than my old 4790k. I think that will stay with me much longer now than anticipated. I expected it to be crushed by the AMD but as it turns out... *shrug*
@phaedarus I think you might as well look at other software to see if that makes a difference. Affinity Photo should have the tools you need and in my experience is generally faster than PS. Not too familiar with the very latest PS versions though. Also perhaps the file format you are using is just slow to handle and things might work better if converted to application native format?
Yeah they're checked. I have the default 70GB of memory allocated. I believe my RAM is 3200mhz CL18 DDR4 using the MSI MOBO profile to increase that to 3600mhz. I just don't understand why Photoshop causes so much lag on my system that even the application switcher slows down (alt-tab). I have a separate 1TB M.2 NVME drive just for cache. No other app behaves this badly.
Hello,
Do you think it's reasonable to switch from a 3900 to 5950x ? I'll do essentially 2D illustrations, 3D modeling, texturing (SP), baking (SP, toolbag) and Unity/Unreal integration ( with lighting). I can get one for 460€. With AM5 which start in fall, i'm still hesitating to make the step.
In same time, the new architecture would be far more expensive (CPU :500-600€ for 12C/24T- mobo : 130-150€ - 32Go DD5: 250-300€).
For any texture or light baking that uses the CPU it should provide a very noticeable performance improvement.
Game engines like Unreal and Unity will have more consistent framerates due to the better cache design of Zen 3 (Ryzen 5000 series) over Zen 2 (Ryzen 3000 series).
Working in 2D, or standard 3D modeling, are more single threaded tasks that wont really benefit much from it.
Hey guys super random question, I am trying to create a new pc for CG/3D work. I have been reading lots of articles on which processors are the "best" for CG and so these articles have made two categories on which process you should get "active work" and "Render work". I am mainly going to be using this pc for active work for modeling sculpting etc but I also will be rendering on this pc. They are saying to get the intel 12900k for the active work and a threadripper for rendering work. Since the begging I have been looking for a threadripper and was wondering if having the threadripper will lower software performance since it has many cores but they are slow and most softwares use single core. Please let me know on what you guys think I should do I am very lost.
The PC I plan to build, for modelling, texturing, sculpting in Maya, Substance Painter/Designer, Zbrush, Unity and Unreal 4 & 5.
Intel® Core™ i9-12900KF Processor
Asus Z690-Plus TUF Gaming bundkort
2 X Kingston Fury Beast 32GB DDR5-5200 RAM (64gb)
Western Digital Black 1TB SN750SE PCle 4.0 NVMe™ SSD
Kingston NV1 2TB M.2 NVMe SSD
Asus GeForce® RTX 3080 Ti 12GB TUF
Any tips or obvious drawbacks?
Would it be worth sacrificing some components, like less or slower RAM, removing a SSD etc to upgrade from the RTX 3080 TI 12GB to the RTX 3090 TUF 24GB?
Quick question: will a laptop with a GTX 560M be able to run Maya? Rest of the specs are some i7 CPU and 14GB of RAM.
I'm starting a Maya training course and it's the only laptop I have, so I was wondering if I'll be able to use it, at least in the beginning.
Install maya and take a look. If you see something on screen it should be fine as a starting point.
It should be fine for standard modeling and animation tasks, just don't expect to be doing any rendering with it (maybe look into offloading any renders to a cloud service). Also make sure to install an SSD in it if it doesn't already have one.