That's also very true.
Because the D700 is slower than a 7970 (yet has the same feature set), it's pretty safe to say the D700 is not 10x faster. Not even remotely at all close.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Wanting to get a nMP soon. when is the 7,1 coming out?
- Thread starter Korican100
- Start date
- Sort by reaction score
That's also very true.
Because the D700 is slower than a 7970 (yet has the same feature set), it's pretty safe to say the D700 is not 10x faster. Not even remotely at all close.
What? Of course you're not working on a final quality render image in real time. However whatever you have in your viewport is seriously taxing your system as the scene gets more and more complex. Any designer/modeler/animator is going to want to work with the best viewport performance possible.
I don't think anyone's saying they can't offer good performance, but workstation cards regularly perform better in this area.
That is slowly shifting as GPU rendering gets more popular. But you're right, and it's all the more reason to have a GPu that offers the best viewport experience possible.
Ok, I will try to explain this for non 3d people.
In every 3D editing software the polygons/textures displayed by the GPU on the viewport(via OpenGL) are not loaded from disk, they have to be processed by the CPU first, and unfortunately this task is mostly single threaded. Therefore machine with higher single core turbo will have an edge above slower clocked machines(even if those may excels in other multithreaded tasks).
For this reason you will see the same identical GPU getting very different results just because they are feeded by different CPUs. And being the old macpro very slow for single threaded operations, you will get a clue of why a 980 will get 50FPS on an old mp and 190FPS on an overclocked modern i7.
Honestly I do not know exactly the difference between a D700 against a 980/oMP, but recently I was surprised to discover that in a real world scene the D700 performs as fast as a 980 on an overclocked PC, and almost 3x faster when activating antialiasing/shadows in the viewport(this is likely due to Firepro drivers optimization). Now if the D700 can be up to 3x faster than a 980 on a fast PC you can just imagine how much faster it will be against an old (5 years technology) macpro.
I'm a 3d guy and I'm mostly concerned by the GPU viewport speed for my works, this is what really matters for me. The 980 should be faster in other tasks like OpenCL and obviously CUDA(don't know about gaming.. I'm not into that stuff).
Last edited:
The specific reason I picked the OpenCL benchmarks is because none of what you're saying is relevant to that.
Most pure synthetic OpenGL benchmarks cache everything they can to VRAM, but OpenCL is more cleanly separated from the CPU. The data doesn't need to be "primed" as much or at all.
I don't know of many people priming OpenGL data with a single thread any more. You certainly don't have to, and if you're doing that, you're doing it wrong. But regardless, that's why these tests are all being done on the same box.
I don't know where do you freelance, but i worked with all the big 10 in EU, US and Israel and i don't see anything from what you mention. I have my own company by the way, so i have a LOT more clients then just visiting 1 studio/year(i had Allianz, Erste Bank, Tuborg, Shell etc in client pool, just to know what we are talking about here right?). Linux is only used for render farms, and OSX is less then 3-5% for 3D work. in add agencies has decreased from 30-40% to less then 15%. Vast majority is Windows for work and Linux for render and network managers..... Again, print, video and advertising. I don't know ANY major production house(or advertising agencies) that switched to Windows who has returned to Apple. Care to provide one example and if i worked with them i will ask and come back to you with facts.
I never once said it was 10x faster. Get your fact straight there bud. I said there are a lot of reasons that a pro card is different to a games card.
----------
So every single thing you have said is about THINGS IN THE FUTURE. What's that got to do with the discussion in hand and the hardware available now.
Given this is a discussion about Pro kit not a gaming PC. What I do have very relevance, not that it had anything to do with you anyway.
I quoted you. You can't even edit your history after I did that. Just admit it was a silly thing to say. No biggie. A Dual D700 set up performs great but there is no need to exaggerate.
----------
I'm seeing hyperbole but no benchmarks or videos here. When you make a comparison it has to be single GPU vs single GPU and SLI vs SLI. Be precise and specific about the sets ups you witness otherwise it's not worth a discussion.
Nope. Can't admit anything as you were quoting out of context Siro76 in his post 271
I have no idea what he is talking about there.
I made not such claim. My entire argument has been that Certification in Apps is important in a work environment and When I have used ( and still ) do use gaming cards, there are often glitches. Sometimes this is not important, but in Building work it's vital. Pro cards are often better at handing large scenes, And games cards are normally faster and Textures.
Lol. My point is that nothing you've mentioned is unique to a pro graphics card and is readily available in consumer cards. All you are paying for is driver signage and in some cases extended OpenGL libraries. You need to stop with the smug omg I'm a pro attitude. You need to let it go
[/COLOR]
And this is what you are still not getting. Unless you have a certified Card... you don't get support.
Actually, this is the most pointless pissing contest.
1. We have a D700 in the nMPro and the other slightly pointless ones
2. Yes there are faster cards out there now - Gaming and Pro.
3. 90% of people can use a gaming card and be perfectly happy. Even in a pro environment.
4. Games cards are cheaper than Pro cards by a factor.
5. If you have to have Cuda - Get a PC and Nvidia
6. If you really dislike windows - Get a Mac.
Done.
And this is what you are still not getting. Unless you have a certified Card... you don't get support.
Actually, this is the most pointless pissing contest.
1. We have a D700 in the nMPro and the other slightly pointless ones
2. Yes there are faster cards out there now - Gaming and Pro.
3. 90% of people can use a gaming card and be perfectly happy. Even in a pro environment.
4. Games cards are cheaper than Pro cards by a factor.
5. If you have to have Cuda - Get a PC and Nvidia
6. If you really dislike windows - Get a Mac.
Done.
I must make the apology that I am confusing you with Siro as you are both overlapping and it is difficult to note user names on my iPhone screen.
I know what thee benefits of a certified workstation card are, as I state my first such card was in the nineties. But to be frank the features they are said to be better at are very minor and don't reflect much difference in the real world. Certainly not the 3 - 10 times difference cited by Siro. The only time a 7970 beats a GTX 980 is in apps that NVidia doesn't want to optimise for. In most cases the 980 is very far ahead of the 7970. And the D700 has slower single GPU performance than the 7970 even because it's a low power custom version. It shouldn't be compared to any new card.
As Intel are starting to wind down production of Ivy Bridge Xeon's I'd say not long now. My sweepstake of an announcement in the last week of January is still on, though I am still waiting for more info to confirm some more things to be absolutely certain.
Without a Xeon and its chipset you have no Mac Pro regardless of what GPU cards are in the 7,1.
Without a Xeon and its chipset you have no Mac Pro regardless of what GPU cards are in the 7,1.
Going with the CPU/chipset releases is definitely the safest guide as it is the main factor - that is, if anyone can guess Apple's plans (e.g. I wouldn't be shocked if Apple chose to skip a Xeon iteration, or release a new revision of nMP with a couple of months delay).
A correction to the thread's title, though, could be that if there's a new release of nMP, it should be the 6.2. An actual 7.1 release should be far, far away yet (if ever).
So, to try and answer the thread's title; I can't tell what the CPU of the 7.1 will be or when this is going to be released, but I predict huge quantities of glue inside
Guys, you look like 10 year olds... :-(
Gav Mack, let's hope you are right. Indeed, Intel will want to let go of IB-E production in favor of Haswell-E.
That might convince Apple to go for it.
Thing is, what GPU will they use? Current Tonga/Hawaii? Wait for Caribbean Islands? More likely the first I guess, if the update is to come sooner than later.
Gav Mack, let's hope you are right. Indeed, Intel will want to let go of IB-E production in favor of Haswell-E.
That might convince Apple to go for it.
Thing is, what GPU will they use? Current Tonga/Hawaii? Wait for Caribbean Islands? More likely the first I guess, if the update is to come sooner than later.
So, to try and answer the thread's title; I can't tell what the CPU of the 7.1 will be or when this is going to be released, but I predict huge quantities of glue inside
Well, I know this is partially a joke, but I wouldn't expect any significant case design changes.
Well I finally bought the base 4c model this week on sale ($2,599) so that can only mean one thing: The new-new mac pro will drop within a month.
I'd like to say that is a joke but it usually happens to me each major purchase. But good news for everyone else!
Going with the CPU/chipset releases is definitely the safest guide as it is the main factor - that is, if anyone can guess Apple's plans (e.g. I wouldn't be shocked if Apple chose to skip a Xeon iteration, or release a new revision of nMP with a couple of months delay).
A correction to the thread's title, though, could be that if there's a new release of nMP, it should be the 6.2. An actual 7.1 release should be far, far away yet (if ever).
So, to try and answer the thread's title; I can't tell what the CPU of the 7.1 will be or when this is going to be released, but I predict huge quantities of glue inside
Going with the chipset usually is the safest bet - only the cMP was pretty much the exception to this rule but that was likely because the 2 piece logic board would have required an entire redesign with no northbridge on the sandy bridge e chipset and the memory controller on the cpu.
The 1-4/5,1 have all pretty much shadowed Intel chipsets in pretty much the same enclosure and I don't expect the Haswell Xeon version to be any different.
----------
Well I finally bought the base 4c model this week on sale ($2,599) so that can only mean one thing: The new-new mac pro will drop within a month.
I'd like to say that is a joke but it usually happens to me each major purchase. But good news for everyone else!
We call it Sod's law over here in the UK! And you did get the base model after all. If it fits your uses don't worry you could always drop an 8 core 3.3 when they drop in price used!
Here it's Murphy's I believe
Unfortunately budget didn't allow for what I wanted originally (6c / 512GB / D500) because of some last minute expenses. My wife needed a new computer so my cMP goes to her and the nMP to me. Merry Christmas everyone.
AD, good deal there.
You guys are so lucky over there, prices are way lower.
Over here the base 4c goes for 3099,00 - that's EUR :-(
Playing the waiting game can be pointless, if the 7,1 is a no show.
If your wife didn't mind you getting the gorgeous new MP while she keeps the old one, lucky you!!
Merry Xmas you too.
You guys are so lucky over there, prices are way lower.
Over here the base 4c goes for 3099,00 - that's EUR :-(
Playing the waiting game can be pointless, if the 7,1 is a no show.
If your wife didn't mind you getting the gorgeous new MP while she keeps the old one, lucky you!!
Merry Xmas you too.