Not necessarily. The term "prosumer" has been a descriptor for both the consumer and product line for quite some time. I'd say it's usually more geared towards the product, especially within the film and video realm.
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.
Will there be a new Mac Pro in 2015?
- Thread starter Anto38x
- Start date
- Sort by reaction score
- Status
Not necessarily. The term "prosumer" has been a descriptor for both the consumer and product line for quite some time. I'd say it's usually more geared towards the product, especially within the film and video realm.
Good points!
Just riffing a bit on the core count... if Apple was to adopt this new Intel® Xeon® Processor E3-1500M v5 Product Family for mobile workstations, which in Intel's own words is targeted towards "content creators, designers and engineers," then one way they could handle it is this:
Being a mobile chip, presumably the TDP is designed to be relatively low. So the nMP's CPU side of the triangular heat sink should have no trouble cooling two of them. Apple could then revert to the old Mac Pro strategy of having an optional two CPUs in the machine.
So if these E3 Xeons were available in 4, 6 or 8 core flavors, doubling them could give CTO options of 8, 12, or even 16 cores.
Core count bragging rights maintained!
Regarding the use (or lack thereof) of GPGPU compute in mainstream apps, possibly this year's WWDC State of the Union feature on Adobe Illustrator using Metal for more instantaneous rendering was intended to nudge developers in that direction?
Anyway, I'm sure I'm getting carried away seeing hot Mac Pro keywords like "Xeon," "Skylake," "Thunderbolt 3," "digital design," "workstation" all being used in the same Intel press release for things shipping imminently. I just want this stuff to come to the nMP sooner rather than later.
Let's wait for the 1500 series specs to come out. If they're somewhat close to the 1600's this could be a solution for some.
I imagine the PCH will be new, in line with what the current Z170 series for the desktop provides.
Enough PCIe 3 lanes for Alpine Ridge, DMI3, maybe Intel even went as far as upgrading USB to v3.1, who knows.
That would be something to write home about.
I bet performance would be lower, some capping must have taken place, not just the core count I'd say.
If it goes up to 8 cores it's good enough for me.
On the other hand, I wouldn't count much on a dual CPU option, too much rework to be done, but who knows.
The option od a dual CPU+ one GPU, or dual GPU+ one CPU would certainly bring peace to a few
But being a 1x00 series it's definitely a 1S CPU, so no way.
Intel also has the E3-1200 but Apple never seemed enthusiastic about it. Or even the Xeon-D. Intel is becoming too diversified in this area.
Maybe the 1500 are more inline with the "normal" Xeons. Maybe it becomes the 1S option for workstations and the 2S+ Purley platform will go for the really high end.
This could be in fact what Apple has been waiting for. It would (probably) bring the needed lanes for TB3. DDR4 with ECC is apparently supported, let's see if 2400 or higher. With the GPUs in place (not starting again on which) and the 5K TBD and we're good to go.
Still the mobile part of it makes me wonder if this has a place in nMP at all.
Maybe the E5-1600 will come shortly after.
Off topic: been to a great concert just around the corner (www.vagosopenair.eu), great European bands from the old times still rocking hard, no head banging for a long time. Sharon (Within Temptation) - looking awesome as usual, love ya.
I imagine the PCH will be new, in line with what the current Z170 series for the desktop provides.
Enough PCIe 3 lanes for Alpine Ridge, DMI3, maybe Intel even went as far as upgrading USB to v3.1, who knows.
That would be something to write home about.
I bet performance would be lower, some capping must have taken place, not just the core count I'd say.
If it goes up to 8 cores it's good enough for me.
On the other hand, I wouldn't count much on a dual CPU option, too much rework to be done, but who knows.
The option od a dual CPU+ one GPU, or dual GPU+ one CPU would certainly bring peace to a few
But being a 1x00 series it's definitely a 1S CPU, so no way.
Intel also has the E3-1200 but Apple never seemed enthusiastic about it. Or even the Xeon-D. Intel is becoming too diversified in this area.
Maybe the 1500 are more inline with the "normal" Xeons. Maybe it becomes the 1S option for workstations and the 2S+ Purley platform will go for the really high end.
This could be in fact what Apple has been waiting for. It would (probably) bring the needed lanes for TB3. DDR4 with ECC is apparently supported, let's see if 2400 or higher. With the GPUs in place (not starting again on which) and the 5K TBD and we're good to go.
Still the mobile part of it makes me wonder if this has a place in nMP at all.
Maybe the E5-1600 will come shortly after.
Off topic: been to a great concert just around the corner (www.vagosopenair.eu), great European bands from the old times still rocking hard, no head banging for a long time. Sharon (Within Temptation) - looking awesome as usual, love ya.
Impossible as a single computer. There is alot of information packed into Intel Processor names. All the more so the Xeon.
http://www.intel.com/content/www/us/en/processors/processor-numbers-data-center.html
A product named Xeon Ex-1xxxx can't possibly be used in two socket set ups. That first of the four digit subsequence is the number of sockets the processor is designed for. A '1' means a single socket maximum allowed.
Could put two computers inside the same container ( each 1 socket operating independently ), but that really isn't a Mac Pro anymore.
The Xeon E3 product line up takes it baseline design from the mainstream processors in the Core iX family. That family has been capped at 4 cores for last several years and that is extremely likely to continue going forward. The vast majority of the increased transistor budget is going toward more GPGPU cores, not x96 one in that product area. In laptops solutions, all the more so.
This E3-1xxxxM is extremely likely just one of the performance class laptop designs ( 'H' series ) with the E3 differentiating features enabled. Just because a "content creator , design" doesn't mean don't use laptops. This is more about bringing something to market for the workstation laptop subclass of products. It will run hotter and incrementally heavier than the "lightest", coolest laptops. That doesn't necessarily put it in modern desktop range. More so in the range of a 2016 laptop that can do what a 2009-2010 desktop could do.
Nope, Xeon E3 just drags the Mac Pro more firmly back into the same zone the iMac ( and laptops ) operate in.
Last edited:
This has almost nothing directly to do with a Mac Pro.
The real item related to the Mac Pro in the Intel blog is far more so the link to the IDC market analysis used to motive the new mobile product being an enabler for a growth market. Engadget's article had a direct link.
http://www.engadget.com/2015/08/08/intel-xeon-for-laptops/
http://www.idc.com/getdoc.jsp?containerId=prUS25821315
Namely the Workstation market is in the sub 1M/quarter zone and not particularly getting any bigger than that. There is some shift in that market from desktop to laptop. ( Totally not surprising since the same thing has happened in the overall PC market). Same issue; what used to be a "desktop" only class problem workload isn't anymore.
That >4 core is a possibility in the same zone as an asteroid strike directly on the Apple HQ.
v5 ( desktop Gen 6) CPUs being necessary for TB v3 is a myth. The Mac Pro could have TB v3 with current generation Xeon E5 products. The number of sockets probably would not be 6, but 2 is quite doable.
But the iMac doesn't have laptop CPUs packages. Hasn't for several years. This is trendline with about zero motivation based on what Apple has recently been doing.
The Mac Pro does need to differentiate itself from the iMac. Using the same baseline processor design as the iMac's is going to make that extremely hard. First, Just a GPU card (or two) isn't going to be enough. "Embarrassing parallel" workloads that have double digit core count needs are drift GPGPU way (especially when in A/V data. ), but some things are still in the >4 and <10 zone with a mix of scalar and highly parallel load.
Second, the upper end iMacs have two GPGPUs too. They are not symmetrical GPUs but there are two. Drag the x86 cores back into parity and have muted the differences.
Last edited:
The term prosumer cameras is well established in the land of the interwebz. However as apple claims, current mac pro is all pro.
Anyhow... Been giving this discussion a lot of thought. I want a new computer to move from my iMac 2011 with SSD. I need a faster and more CPU powerful machine because I use lots of softsynths.
Am I crazy to be worried about apples future when I see them coming out with watches, TV appliances and undervolt cards?
If I didn't have a thunderbolt interface i would jump to PC now. Damn you UAD
Anyhow... Been giving this discussion a lot of thought. I want a new computer to move from my iMac 2011 with SSD. I need a faster and more CPU powerful machine because I use lots of softsynths.
Am I crazy to be worried about apples future when I see them coming out with watches, TV appliances and undervolt cards?
If I didn't have a thunderbolt interface i would jump to PC now. Damn you UAD
With Skylake coming to market now, the current nMP's Xeon is 3 architecture generations behind the consumer world. Within the same architecture, the following differences usually are mentioned as reason to use Xeons in a workstation environment (please correct me if I forgot something important):
ad 2.: Probably the biggest reason people argue with when it comes to (semi-)professional work. However, since ECC has been introduced, technology (especially production technology) has seen massive improvements, so I wonder if that really is an issue today with modern non-ECC Ram. Are there any studies or other data available that show statistically relevant data that non-ECC Ram is still prone to bit errors? And with resolutions of edited video massively increasing, are rare bit errors still noticeable? I do understand that scientific applications may still be fragile here, but then it's back to my initial question of the quality of today's non-ECC ram in general.
ad 3.: As Skylake i7's are (artificially!) limited to 4 cores, the Xeons seem to have a distinct advantage here. However, Apple seems to think that even 4 cores are a viable option in a workstation. And with architectural improvements 4 Skylake cores may be able to equal 6 or even 8 Ivy Bridge (Xeon) cores. Similar to the old MP 1,1: Even upgraded to 8 cores, a modern mini (at least the 2012's) can beat it.
ad 4.: As Apple has taken a design decision to stick with only 1 CPU chip in the nMP, this advantage definitely does not apply anymore.
To my layman's eyes, consumer hardware is increasingly capable of performing well in tasks where you used to need professional hardware in the past. And with the professional (CPU) platform lagging behind in development (and seemingly with the gap increasing rather than decreasing), I wonder for how long Apple is continuing to put Xeons in the MacPro.
They did away with the Xserve as they didn't see a big enough target group compared to the required effort (certain target groups do expect quite some support). With the design change from the cMP to the nMP they also did away with some parts of the professional target group they did cater for before. The nMP in its current incarnation is restricted in some areas, compared to professional servers, so number-crunchers or scientific users will probably decide for a competitor product anyway, be it because they need more than 128GB Ram or more than 1 Xeon for the sheer core count.
Does the remaining target group (e.g. A/V people) really _need_ Xeons for their work? From what I read here, a significant number is even using iMacs by now.
From my point of view, it would only be consequent if Apple replaced the Xeons at least in the entry-level nMP's with modern i7's (and in the long run in all nMP's, once GPGPU has caught sufficient traction in the professional world). Less so because of (non-existing) price advantages, but for being able to offer up-to-date technology in their flagship product.
Chances are that Xeons will fall behind even further in the coming years, simply because the target group is growing smaller and Intel is focusing on the money-makers (i.e. high-volume products) in their development spending. Apple is receiving flak for not upgrading their flagship product, which they could avoid by switching to consumer CPU's, similar to what they did with the x86 switch, when PPC development fell behind more and more. Only that the switch to consumer CPU's this time would be quite easy on the software side.
They probably would have to optimize the whole package to offer a more competitive product once they can't argue with "professional grade components" anymore (which they arguably can't do anymore even today), such as e.g. slower SSD's or "only" DDR3 Ram or a plastic housing instead of the polished metal they currently ship. But it's not an impossible or even difficult task.
To summarize it: With 3 architecture generations difference, is a Xeon still a viable alternative to a conventional i7 system outside the high-end professional segment, which is not satisfied with current nMP's anyway?
- L3 Cache
- ECC Ram support
- More cores
- Multi-CPU
ad 2.: Probably the biggest reason people argue with when it comes to (semi-)professional work. However, since ECC has been introduced, technology (especially production technology) has seen massive improvements, so I wonder if that really is an issue today with modern non-ECC Ram. Are there any studies or other data available that show statistically relevant data that non-ECC Ram is still prone to bit errors? And with resolutions of edited video massively increasing, are rare bit errors still noticeable? I do understand that scientific applications may still be fragile here, but then it's back to my initial question of the quality of today's non-ECC ram in general.
ad 3.: As Skylake i7's are (artificially!) limited to 4 cores, the Xeons seem to have a distinct advantage here. However, Apple seems to think that even 4 cores are a viable option in a workstation. And with architectural improvements 4 Skylake cores may be able to equal 6 or even 8 Ivy Bridge (Xeon) cores. Similar to the old MP 1,1: Even upgraded to 8 cores, a modern mini (at least the 2012's) can beat it.
ad 4.: As Apple has taken a design decision to stick with only 1 CPU chip in the nMP, this advantage definitely does not apply anymore.
To my layman's eyes, consumer hardware is increasingly capable of performing well in tasks where you used to need professional hardware in the past. And with the professional (CPU) platform lagging behind in development (and seemingly with the gap increasing rather than decreasing), I wonder for how long Apple is continuing to put Xeons in the MacPro.
They did away with the Xserve as they didn't see a big enough target group compared to the required effort (certain target groups do expect quite some support). With the design change from the cMP to the nMP they also did away with some parts of the professional target group they did cater for before. The nMP in its current incarnation is restricted in some areas, compared to professional servers, so number-crunchers or scientific users will probably decide for a competitor product anyway, be it because they need more than 128GB Ram or more than 1 Xeon for the sheer core count.
Does the remaining target group (e.g. A/V people) really _need_ Xeons for their work? From what I read here, a significant number is even using iMacs by now.
From my point of view, it would only be consequent if Apple replaced the Xeons at least in the entry-level nMP's with modern i7's (and in the long run in all nMP's, once GPGPU has caught sufficient traction in the professional world). Less so because of (non-existing) price advantages, but for being able to offer up-to-date technology in their flagship product.
Chances are that Xeons will fall behind even further in the coming years, simply because the target group is growing smaller and Intel is focusing on the money-makers (i.e. high-volume products) in their development spending. Apple is receiving flak for not upgrading their flagship product, which they could avoid by switching to consumer CPU's, similar to what they did with the x86 switch, when PPC development fell behind more and more. Only that the switch to consumer CPU's this time would be quite easy on the software side.
They probably would have to optimize the whole package to offer a more competitive product once they can't argue with "professional grade components" anymore (which they arguably can't do anymore even today), such as e.g. slower SSD's or "only" DDR3 Ram or a plastic housing instead of the polished metal they currently ship. But it's not an impossible or even difficult task.
To summarize it: With 3 architecture generations difference, is a Xeon still a viable alternative to a conventional i7 system outside the high-end professional segment, which is not satisfied with current nMP's anyway?
Single CPU workstations don't need Xeon or ECC RAM if they are used for multimedia production. If your are doing computationally intensive scientific modelling, climate science, virtual animal models for medical research. Etc then maybe ECC is worthwhile. If Apple is going to continue with the Mac Pro brand it needs to expand the range to sell a gaming model ("pro" level gaming rig), multimedia model, an engineering model, and a science model. The first two don't need Xeon. The third can be single Xeon. The last can be dual Xeon.
But my feeling as I have said before is OSX will be released to install on any PC but Apple will tell all those new official converts that their machines are better than the home built boxes people install OSX on. Maybe they will keep some special features just for their own 'real' Macs.
But my feeling as I have said before is OSX will be released to install on any PC but Apple will tell all those new official converts that their machines are better than the home built boxes people install OSX on. Maybe they will keep some special features just for their own 'real' Macs.
What? Seriously, how hard is it to not appear so ignorant when you are clearly connected to the internet?
I like your thought process but apple is unlikely to do this. Apple has hinted interest in the gaming comunity. We can't expect tripped out PC with water cooling performance but there is room for improvement.
It is also possible that apple has a gaming console in the works. Might then over take the apple TV as it stands.
Regarding console I think a future Apple TV will be compatible with iOS games and use a touch screen TV and wireless controllers. The foundation is already laid.
The quality of non-ECC RAM has improved, but only in speed and performance. I haven't see any evidence to show that recent standard memory has fewer bit errors. In the meantime, ECC memory has improved in performance as well.
Also, the failure rate of standard RAM (.6 %) is still much worse than ECC RAM (.09%). There are good reasons why servers and workstations still use ECC registered RAM.
http://www.techspot.com/article/845-ddr3-ram-vs-ecc-memory/
https://www.pugetsystems.com/labs/articles/Advantages-of-ECC-Memory-520/
I maintain a couple of hundred server systems (all HP ProLiants). All with HP memory, no third party crap. (HP memory has some additional error/mgmt. bits)
The servers send me an email when they have an ECC-corrected (single bit) memory error, and an email if they force a machine check (kernel panic) because of a non-correctible multi-bit error. I see a single-bit message every couple of months (over several hundred machines), and panic messages when a DIMM has a catastrophic failure (a few per year).
The advantage of ECC memory isn't that it is more reliable than non-ECC memory - its that when a memory error happens you get a message that "there's been a memory error". With non-ECC memory, you have a strange crash or strange data - without any indication that a hardware problem is the cause.
Not a big deal for the wedding video crowd, but a big deal for many who care about their data.
ps: my home PC is an E5-1650v2 with 128GiB of 1866 MHz ECC RAM. Basically a six-core MP6,1 with more RAM and PCIe slots.
ADK Pro Audio sells PC DAWs with Thunderbolt 2 on the logic board. Would it work with UAD gear?
I'm with you there and glad I used RampageDev to help me with my install. That's why I'm not wasting my time tinkering around with installing things that will only frustrate me. I just like building the systems and either OC'ing or Underclocking them to get the most out my Mac OS X.
Well that wasn't Apple's initial thought process before. They were all about speed and performance. Just to let you know, this is more up my alley:
That other link you gave me to see, didn't impress me much. In fact I'm building something very similar to this that eventually I will be posting soon. Hopefully Mac OS X will be working soon on my new Dual CPU & 2 x 980Ti setup.
You cited that there were 100's of thousands of hackintoshs being built. Any quantitative justification for that number or just pulling numbers out of the air?
Apple is selling around 18-20M Macs per year. Even a hacintosh rate of 200K/yr is about 1-1.1% piracy rate which isn't that bad. It is never going to be zero. Are there more hackintoshes now than 6 years ago? Yes. Are there more Macs sold per year than 6 years ago? Yes.
If the hackintosh guesstimate is being measured by forum/site membership or software download numbers then that is rather dubious. It does not directly measure daily usage of actual systems. There are more than few folks who join sites for other reasons than the primary purpose. Likewise, there are probably more than a few abandoned Frankenstein projects in people's hobby workbenches.
Last edited:
There are some interesting rumors out there today about a Nvidia GTX 990M, which would be a cut down mobile variant of GM200 (aka GTX 980 Ti/Titan X). The TDP ranges from 100-185 W, which would be on the upper end of what could fit thermally. Rumored timeframe is Q4, leaning more towards september. Probably unlikely that Apple goes back to Nvidia for the Mac Pro, but maybe the efficiency of Maxwell is too hard to resist. I could see a lineup of
D310: GM204 cut down, 4 GB VRAM (aka GTX 970)
D510: GM204 full, 8 GB VRAM (aka GTX 980)
D710: GM200 cut down, 12 GB VRAM (aka GTX 980 Ti)
Some nice, clear distinctions in this lineup in terms of performance and VRAM. Apple may be in an uncomfortable position with AMD's recent Fiji chip, which is limited to 4 GB VRAM. They would have to choose between slower Hawaii with more VRAM, or faster Fiji with less VRAM. Apple likes to distinguish upgrade options using VRAM to correlate to increased performance. Going back to Nvidia gives them this. Just because Apple has been going all AMD in recent years doesn't mean they won't switch back and forth, depending on various competitive factors.
Perhaps some supporting evidence might be the big gains in performance for Maxwell seen in the most recent driver updates, measured by Barefeats.
D310: GM204 cut down, 4 GB VRAM (aka GTX 970)
D510: GM204 full, 8 GB VRAM (aka GTX 980)
D710: GM200 cut down, 12 GB VRAM (aka GTX 980 Ti)
Some nice, clear distinctions in this lineup in terms of performance and VRAM. Apple may be in an uncomfortable position with AMD's recent Fiji chip, which is limited to 4 GB VRAM. They would have to choose between slower Hawaii with more VRAM, or faster Fiji with less VRAM. Apple likes to distinguish upgrade options using VRAM to correlate to increased performance. Going back to Nvidia gives them this. Just because Apple has been going all AMD in recent years doesn't mean they won't switch back and forth, depending on various competitive factors.
Perhaps some supporting evidence might be the big gains in performance for Maxwell seen in the most recent driver updates, measured by Barefeats.
There is no indication in El Capitan beta for Maxwell drivers. Also in Barefeats tests we see increases from Nvidia web drivers.
There is a catch in those tests. In games GTX 680 is as fast as GTX 980 Ti. Which in real world is BS. GTX 680 is slower than GTX 980 and what about 980 Ti?
P.S. I have a feeling that that Maxwell cut down mobile chip is an answer to AMD mobile proposition...
There is a catch in those tests. In games GTX 680 is as fast as GTX 980 Ti. Which in real world is BS. GTX 680 is slower than GTX 980 and what about 980 Ti?
P.S. I have a feeling that that Maxwell cut down mobile chip is an answer to AMD mobile proposition...
That's an insane modular design
What clock speed do you run your SR2 at? Is it air or water cooled?
Did you see the benchmarks I posted? Even with the new drivers OSX is so far behind Windows it is sad and funny at the same time.
- Status