Will there be a new Mac Pro in 2015?

[deconstruct60]

I agree that there isn't necessarily a need for 3 thunderbolt controllers and 6 USB-C ports. I think the rationale for 3 controllers and 6 ports with shared bandwidth is simply because the ports share double duty with displayport. On the current mac pro, if you want to run 3 displayport monitors, that leaves you with 3 TB ports and 2 TB controllers for peripherals.

Sigh.... how thunderbolt does and doesn't work has been covered multiple times in this thread. You can saturate the PCIe feed to a TB controller by using just one of the two ports that a single controller provisions. What you are implicitly assuming is that users will cover two TB ports from a single controller with two DisplayPort (and/or two USB 3.0 Type C for TB v3). They don't have to. If tip-toe just the right way and take measure of how everyting is hooked up inside. But that is not particularly plug-and-play Apple style. Stretching to three 4K monitors you have to do a bit of that now, but adding in the additional multiple dimensions of the shared USB 3.0 controllers inside the TB controller and displayport and oversubscribed PLX switch. You shouldn't have to look at an internal circuit diagram to pick out what port to plug into. That throws "one plug to rule them all" out the windows when it comes to improving simplicity.

However, I don't see Apple offering USB-C and miniDP on the same machine. I can't think of a case where Apple has increased the variety of ports.

MBP has miniDisplayPort and HDMI when used to just have whatever single video solution Apple was rolling with before Thunderbolt came along. Thunderbolt has been the "more than one" option since it appeared.

What I don't see Apple doing short term is "plain" USB 3.0 Type C and Thunderbolt v3 flavored Type C on the same system. It is the same physical plug but they have different ranges. Apple has avoid that. They went all USB 3.0 instead of mixing matching USB 2.0 / USB 3.0 with different colors.

Apple has also gone FW400 and FW800 for a while on a single system during the Firewire's evolution.

2008 Mac Pro https://support.apple.com/kb/SP11?viewlocale=en_US&locale=en_US

2007 MBP https://support.apple.com/kb/SP13?viewlocale=en_US&locale=en_US


It would not be unusual or unprecedented at all. It is not going to buy Apple many friend to have those limited set of folks who did may the TB v1-v2 investment to have to run out and buy even more dongles. Honestly it is a bit pretty. Even more so when the Xeon E5 v4 + 612 chipset isn't quite ready to an increase of both SSD and TB v3 at the same time. The timing is off to do both at the same time.

Since we already have USB-C on the macbook, I foresee USB-C becoming standard across the line, with everything above the macbook getting USB-C/TB3, replacing miniDP and the magsafe connector.

There is zero rational need for that where there actually is space on the connector panel edge. Dumping all of the 1-2 USB ports into Type-C. Sure. That makes sense because there are just room for those two. But when there are 10+ ports why trying to push them into just one type. With limited space there is added value in doing more with less. With lots of available space it is just dogma. There is no basic need. Space is available and doable with dramatically less resistance from folks about having to buy dongles.

There no good reason to replace magsafe on a 15" or 13" Macbook. The only reason the Macbook goes down to one is the depth is so short which is driven by the sub 13" screen coupled to maximum thinness. (keeping the physical audio jack is another minor point). A very reasonable upper/lower bezel addition and it would work ( the weight savings ouldn't be quite as high though) With a 13" screen (since screen is taller and hence system has to be deeper) there room for 2-3 ports of which magsafe is one. A super light laptop should drop the utility of Magsafe because the laptop is so light it won't detach quickly enough. For a device the size of the Mac Pro it is in the utterly ridiculous zone that have to maximize socket conservation down to just a handful of ports.

 

[jonisign]

Alright, I answered my own question, it turns out that Thunderbolt 2 is connected via the PCH almost all PC's. This differs from Apple's implementation where it is hung from the CPU lanes. This suggests that Apple could put the TB 2 on the PCH, and TB 3 on the CPU. So, my current prediction is

2x TB 3 ports (x4 PCIe 3.0 lanes, CPU)
2x TB 2 ports (x4 PCIe 2.0 lanes, PCH),
and SSD support (x4 PCIe 3.0 lanes, CPU)

 

[MacVidCards]

http://www.golem.de/news/grafikkarte-amd-benchmarks-sehen-r9-nano-vor-der-r9-290x-1508-115897.html From the view of performance the core clock should be around 850 MHz with 150W of power consumption. The power consumption is calculated however, only from the results, and 260W of power consumption for R9 290X.

Not bad, not great, IMO. Was hoping for more.

Anyways, counting the performance it will be around 7 TFLOPs of compute power from 175W of TDP.
Titan X has 250W of TDP and 6.7 TFLOPs of compute power.

Im starting to wonder how that will affect the Mantle-ish applications, where compute power and asynchronous shaders are the most important factor.

Things are getting interesting.

Looks like another base hit from AMD.

Keep in mind that pre Fury launch they announced that it was "fastest card in the world" and an "overclockers dream". Both of these phrases have been proven to be actual lies.

So don't be too surprised if Nano isn't measurably faster then 390x . The presence of the word "Hawaii" instead of a model number tells me they choose a lower end one like a 290 (no "x")

Wouldn't be so suspicious but their CEO got up on stage and told lies 2 months ago, massaging facts has been their modus operandi before.

And I'm going to go nuts pointing out that 2 @ 175 Watts is still writing a 100 Watt check that nMP can't cash.

 

[SoyCapitanSoyCapitan]

The design will be likely tweaked at some point so no point speaking about the future model in terms of the first one. The power supply capacity could be increased, the cylinder could become taller, or shorter and wider, or oblong instead of circular. It could have a built in security lock. It could become metallic chrome to emphasise Metal. It could have an SD slot.etc

 

[koyoot]

Fastest card is a lie you say?
http://cdn.wccftech.com/wp-content/uploads/2015/05/AMD-Fiji-GPU-Powercolor-Giveaway.jpg Sneak peak before the presentations.
http://cdn.wccftech.com/wp-content/uploads/2015/06/AMD-Radeon-Fury-Series-Fiji-GPU-Cards.jpg

Actual slide on the presentation. It plainly says that Dual Fury X is the fastest GPU in the world. You messed with it a bit. And accused a CEO of other company as being a liar. Hmmm, Im wondering what will you say about a company that offers 3.5 GB GPU as 4 GB(GTX 970 case).

However I agree with you. In the end it may be that Fury Nano will be only slightly faster in DirectX 11 games than R9 390X or at the same level of performance. The problem goes for DX12. 7 TFLOPs of compute power will pay off in comparison to 5.9. Not by high margin, but anyway it will. DX12 and every other Mantle-ish API will bring compute power as the factor of performance, thats why there is not that big difference in games between R9 290X and GTX 980 Ti in DirectX 12. So 7 TFLOPs Fury Nano might end up being faster than Titan X, or on par with it.

P.S. I would wait for the actual figures of power consumption and power from Fury Nano. We can still be wrong, tho. With Mantle AMD pulled the right trigger, lets see what they will do with Nano.

 

[koyoot]

The design will be likely tweaked at some point so no point speaking about the future model in terms of the first one. The power supply capacity could be increased, the cylinder could become taller, or shorter and wider, or oblong instead of circular. It could have a built in security lock. It could become metallic chrome to emphasise Metal. It could have an SD slot.etc

It also can have Water Cooling. Which IMO would be a killer feature, for me at least .

 

[SoyCapitanSoyCapitan]

It also can have Water Cooling. Which IMO is a killer feature, for me at least .

If it's not updated until Skylake Xeons it won't even need water cooling. A quad core Skylake can run safely at 4.8ghz on air and achieve over 30,000 geekbenches on those four cores alone. Imagine what a 12 core Skylake would achieve compared to the Haswell.

 

[koyoot]

Not the CPU would benefit most from Water Cooling .

The GPUs would .

 

[MacVidCards]

Not the CPU would benefit most from Water Cooling .

The GPUs would .

As I was trying to point out before, the second a nMP needs water cooling, it also needs a new PSU. Heat out is a direct function of power in.

So to run some downclocked and down volted Fury Nanos, they would also need an external PSU brick or bigger one inside. Once you add all these cables and units to the TB pile, you have officially created a desk octopus.

Which leaves one wondering WHY it ever needed to be shrunk to a point where it can neither power, nor cool, leading desktop GPUs.

Here come the Mobile GPUs. As I have pointed out, my D300s are already such a thing.

 

[JamesPDX]

If it's not updated until Skylake Xeons it won't even need water cooling. A quad core Skylake can run safely at 4.8ghz on air and achieve over 30,000 geekbenches on those four cores alone. Imagine what a 12 core Skylake would achieve compared to the Haswell.

Do you think Apple will stick to the unified thermal core design? "...it’s now hard to imagine building one any other way."
I'm starting to think that if they stick with this design and just keep making things more powerful, maybe I won't look for a 2010 or 2012 as my next machine. -But I'd still have to go for a Blackmagic Design Multidock 2 for external storage in the studio. A modular computer would be best and I think it will be the future if concerned companies are legit about "The Environment" -look at this as an idea of ongoing utility, getting what you need instead of Choiceless Plus: http://www.burlaudio.com/products/b80-mothership

 

[R3k]

At work I take care of I run 3 x 2010 hex core Mac Pro's similar to your spec and also one 2014 cylinder nMP.

Heavy load music / film score work, 32gb RAM per system, 5 hard drives per system (10tb), 2 x 27" (2560 x 1440) screens along with one 1280 x 720 screen per workstation.

The new Mac Pro is not that much better in perceivable performance, the older machines are still great. I have no plan to upgrade them at this point.

 

[deconstruct60]

It's certainly interesting to see Dell and HP move in so aggressively. ....
I'm wondering how effective a long-term business strategy it can be, though.

Dell and HP own around 30-40% of the workstation market each (so 60-80% of "workstations sold as systems" market total. I think there is a more hard to measure market of folks build limited runs for themselves... same uptick of that in segments of the server market with Google/FB/etc. making their own servers. ). Not as much as aggressive in an expansive way but aggressive in a defensive way. They already sell a substantial number of systems into this "creative" market even before the Mac Pro changes. It is Apple that has a problem of being a smaller player and possibly getting even smaller with respect to the workstation market.

The Mac Pro is probably a "hobby" product for Apple at this point. As long as it sells above of a certain threshold they will keep selling them. If it falls below then they probably find another hobby. if HP and Dell can push the Mac Pro units under that threshold number they kill off a weaker player in the workstation market. Most of of the refugees from that implosion will probably go to them since they are two biggest players. That won't necessarily grow the worksation market larger but those two will survive at least a couple years longer. A very long term solution? No (can't keep killing off the smaller palyers forever if own more than 60% of the market) but at least they make it to the future a couple years down the road and try to respond to that new context.

desktops are increasingly irrelevant for a large part of the public, and workstations themselves following the same trend. For now, they're the only place where there's still a strong technical reason to demand new hardware—I can do all the page layout, spreadsheets and word processing I want on a PowerMac if I wanted to.

A far number of workstations were desk-side or under desk more so than desktops. The current Mac Pro design is clearly aimed at being strictly a desktop (and desktop without consuming a large chunk of the desk surface area/volume either. ). Whether that transition is going to grow fast enough for Apple to stay interested is an open question.

But yes... there are more than a few subgroups whose computation workloads are not growing faster than the hardware+software capabilities are. As long as not in the game of running old legacy software with a bigger hammer and more open to new software that runs more efficiently on newer hardware they aren't as out of alignment as many of the complaints in the forum would indicate.

According to this report, the workstation market is essentially <1 million units. [1] It's a damn small slice of pie to keep fighting over and to rely on heavily. Apple is insulated from their own failings and the larger whims and malaise of the PC market. Will be interesting to see how HP and Dell fare.

I'm not sure how HP services is splitting up in the PC/Printer vs Enterprise/Software break up coming but Dell will still be coupled and private. These "workstations" could be servers for smaller companies and there is alot of overlap in parts for the more dedicated rackmounted servers. As long as Dell and HP have a healthy server biz then having workstations isn't a big leap ( more like having iPads if have iPhones ) than it is for Apple which it is a outer fringe of their product ecosystem.

If Dell and HP could get rid of the bulk of their PC line up that has razor slim and/or no profit they'd be in decent shape. The workstation stuff probably has reasonable margins. The workstation and more general PC line up are more bloated than necessary though. Being everything to everybody isn't going to work when you are just a PC&Printer vendor.

Apple isn't as insulated as you make them out to be. They are leveraging missteps by Windows/PC vendors to grow faster than base market ( largely I think by taking share away from weaker players). They can play on that for a substantial amount of time since starting as a very small player but as extend out in to the long term they will have problems ( can't play the zero sum game forever. )

 

[Kpjoslee]

Fastest card is a lie you say?
http://cdn.wccftech.com/wp-content/uploads/2015/05/AMD-Fiji-GPU-Powercolor-Giveaway.jpg Sneak peak before the presentations.
http://cdn.wccftech.com/wp-content/uploads/2015/06/AMD-Radeon-Fury-Series-Fiji-GPU-Cards.jpg

Actual slide on the presentation. It plainly says that Dual Fury X is the fastest GPU in the world. You messed with it a bit. And accused a CEO of other company as being a liar. Hmmm, Im wondering what will you say about a company that offers 3.5 GB GPU as 4 GB(GTX 970 case).

However I agree with you. In the end it may be that Fury Nano will be only slightly faster in DirectX 11 games than R9 390X or at the same level of performance. The problem goes for DX12. 7 TFLOPs of compute power will pay off in comparison to 5.9. Not by high margin, but anyway it will. DX12 and every other Mantle-ish API will bring compute power as the factor of performance, thats why there is not that big difference in games between R9 290X and GTX 980 Ti in DirectX 12. So 7 TFLOPs Fury Nano might end up being faster than Titan X, or on par with it.

P.S. I would wait for the actual figures of power consumption and power from Fury Nano. We can still be wrong, tho. With Mantle AMD pulled the right trigger, lets see what they will do with Nano.

Not sure why are you expecting Fiji to be in next Mac Pro update....current HBM is limited at 4GB, it will probably be lower clocked 390x equivalent with 8GB of memory if they are going to update.

 

[MacVidCards]

Fastest card is a lie you say?
http://cdn.wccftech.com/wp-content/uploads/2015/05/AMD-Fiji-GPU-Powercolor-Giveaway.jpg Sneak peak before the presentations.
http://cdn.wccftech.com/wp-content/uploads/2015/06/AMD-Radeon-Fury-Series-Fiji-GPU-Cards.jpg

Actual slide on the presentation. It plainly says that Dual Fury X is the fastest GPU in the world. You messed with it a bit. And accused a CEO of other company as being a liar. Hmmm, Im wondering what will you say about a company that offers 3.5 GB GPU as 4 GB(GTX 970 case).

However I agree with you. In the end it may be that Fury Nano will be only slightly faster in DirectX 11 games than R9 390X or at the same level of performance.

AMD gave every impression that they were introducing the fastest GPU in the world.

Their official Reviewers Kit came with a graph that showed Fury-X beating 980Ti in 12 out of 12 tests. As we all know, not a single reviewer since has gotten anything like this.

http://www.forbes.com/sites/jasonev...-new-fiji-graphics-card-beats-nvidias-980-ti/

They also claimed that it beat 980Ti in 3D Mark Ultra Firestrike.

Here is what Forbes Magazine said about that:

"AMD's internal benchmarks show a FireStrike Ultra score of 3817 for the Nvidia 980 Ti. This is using their primary test bench system which is driven by an Intel Core i7-5960x. It's a system almost identical to my own, on which I tested a reference 980 Ti on FireStrike Ultra and received a score of 4007. This is just a reminder to take any internal benchmark results with a grain of salt."

Again, a feat that nobody has been able to repeat. In fact, here are the top 100 Firestrike scores. Only a single entry with AMD by it. The other 99 winners were all Nvidia.

http://www.3dmark.com/hall-of-fame-2/fire+strike+3dmark+score+ultra+preset/version+1.1/1+gpu

And if the way out of the past lies is to say "Oh, the card we have coming several months from now WILL be the fastest", what is that? Only for entry in the Vapourware 500 in my book. Nvidia could say same thing, "Our new Pascal D1000 takes any and all comers, past, present, and future". Claiming you have the fastest GPU on the planet but it is super secret and only testable in your labs....I'm laughing while I type it.

I get that you are on the AMD cheer team. And I would agree that it is possible that we wouldn't have a 980Ti if there wasn't a Fury. But making wild claims like "fastest card" and "overclocker's dream" and "speed of Hawaii with 175 Watt power draw" and "Nano will fit in nMP and still delver Fury-X performance" are what has gotten AMD in the egg-stained-face position it is in now. I am attaching the graphs that AMD provided to the press as a "reviewer's guide" showing Fury-X beating 980Ti in 12/12. Not even slightly credible. Why create bogus graphs and figures when you know reviewers will publish the truth shortly thereafter?

 

[koyoot]

Not sure why are you expecting Fiji to be in next Mac Pro update....current HBM is limited at 4GB, it will probably be lower clocked 390x equivalent with 8GB of memory if they are going to update.

Because of this:

Fury X device ID in ElCap DP1 and DP2 drivers could indicate that some work is in progress...
In DP2 they added additional framebuffer personalities and now AMD9000Controller contains first complete personality "tailored" for the trashcan. Of course still no Fiji section in AMDX4000.

I'm not saying that it will happen, but just pure facts I noticed.

MVC, You completely derailed what was said, turned to your idea what was said by their CEO. Im sorry, but I don't have time to argue with you. Im ignoring you from now on.

Edit: http://semiaccurate.com/forums/showpost.php?p=243026&postcount=1209 Interesting post about OpenCL drivers, and Folding@ Benchmark.

 

[fuchsdh]

If Dell and HP could get rid of the bulk of their PC line up that has razor slim and/or no profit they'd be in decent shape. The workstation stuff probably has reasonable margins. The workstation and more general PC line up are more bloated than necessary though. Being everything to everybody isn't going to work when you are just a PC&Printer vendor.

Apple isn't as insulated as you make them out to be. They are leveraging missteps by Windows/PC vendors to grow faster than base market ( largely I think by taking share away from weaker players). They can play on that for a substantial amount of time since starting as a very small player but as extend out in to the long term they will have problems ( can't play the zero sum game forever. )

Yeah it's amazing how many SKUs Dell and HP offer. You can fault Apple for cutting away so many designs that they don't serve important niches, but even as a tech-literate guy trying to figure out what kind of PC I'd get is an exercise in frustration (helped along by the skeevy "always on sale" prices they have and the weird BTO options.) They have seven consumer desktop models, not counting different configurations and the Alienware line. They have nearly as many in the business section, and that's all before you hit real workstations. I don't understand how that is an effective strategy at all or they're still doing that in 2015.

Thus far, Apple's shown that people are willing to pay a premium for a nice design and nice experience; really that's been the case even when they weren't all that competitive on paper with PCs. I don't particularly see that changing any time soon... Apple has basically the best customers possible—the ones that aren't going to leave them just because they can find a cheaper piece of plastic elsewhere. From a consumer standpoint I don't think they're really in direct competition at this point; there's Apple and a few other vendors/special computers, and then there's all the old PC makers fighting for a decreasing slice of the pie.

 

[koyoot]

The problem is this. Desktop PCs are declining in sales. Laptops are declining in sales. In both Apple sees growth in market share, and in actual sales numbers.

The market on servers and workstations is also growing. To the rate in which Apple may be seeing decrease in market share, but in overall sales - increase.

Lets assume the best possible scenario:
Broadwell - EP CPUs
DDR4 RAM with 2400 MHz clock rate.
Faster and bigger SSDs.
AMD GPUs with GCN 1.1 or higher feature set.
Thunderbolt 3.
Metal low-level API.
New OpenCL drivers.
New API for RayTracing on OpenCL in OSX.
Updated design with better reliability.
Water Cooling.

Market share would go up with all this? Don't think so. But the sales would skyrocket. What I see in Mac Pro is that it is a day-to-day user friendly Workstation, and I believe it was designed for that type of consumer. For people who need really powerful computer, and can afford it. For people who actually earn money from work on computer, and want to use Apple ecosystem, or applications. For people who don't want to tinker with their computer a lot, and just use it.


And I don't think Apple cares about market share if they see in the end boost in sales of their products.

P.S. There is a word from game developer that Fury Nano might be faster than Fury, and in the range of Fury X.
http://forums.anandtech.com/showpost.php?p=37653990&postcount=122
If this is the case it is a killer card, but the problem is the price here.
http://forums.anandtech.com/showpost.php?p=37654029&postcount=128
Well...
Continuation:
http://forums.anandtech.com/showpost.php?p=37654069&postcount=136
http://forums.anandtech.com/showpost.php?p=37654072&postcount=137
So as I suspected. Low voltage with TDP limit. The core clock will jump between numbers. In some cases it will be 850 MHz in some cases it will be 1000 MHz. All depending on the workload, and power envelope.

 

[deconstruct60]

The problem is this. Desktop PCs are declining in sales. Laptops are declining in sales.

Not really.

"... The decline was led by a double-digit decline of desk-based shipments, which offset single-digit growth of mobile PCs. ..."
https://www.macrumors.com/2015/07/09/apple-us-mac-sales-drop-2q15/

There are subsets of the PC market that are doing OK (not super go-go growth of 80-90's but OK). There are other parts that are dying ( or minimally going through a major retrenchment ) . What is flawed is the cover every possible SKU/configuration possible strategy. In a hectic overall high growth market that may appear to work. In a market where there is extreme maturity and/or major user need changes it probably won't.

In both Apple sees growth in market share, and in actual sales numbers.

Because 'better' or because other folks are screwing up?? It is a bit of both. Apple has avoid the downturn segments like the plague. Windows 8 "fear" and major PC vendors restructurings ( HP spun out or not, Dell goes private. IBM steps away from x86 boxes completely after letting go of last part of x86 servers )

The market on servers and workstations is also growing. To the rate in which Apple may be seeing decrease in market share, but in overall sales - increase.

Increase? ROTFLMAO. Workstation is down.

"... the industry shipped approximately 912.4 thousand workstations in Q1'15, equating to a sequential decline of 11.7% and a more modest 3.5% year-over-year decline. .."
http://jonpeddie.com/publications/workstation_report/

"... Worldwide shipments for the quarter recorded a year-over-year decline of -1.7%, less than forecast projected growth of 2.4%. Overall shipments fell to 838,400 compared to 858,667 in the same quarter a year ago. ... "
http://finance.yahoo.com/news/overall-certified-workstation-growth-more-212000499.html

"... Overall worldwide shipments improved slightly with a year-over-year decline of -1.0% compared to a year-over-year decline of -1.6% in the first quarter of 2015. ..."
http://www.idc.com/getdoc.jsp?containerId=prUS25821315

That was all before the major stock market correction. Alot more businesses are going to throw their budgets on "essentials only" lockdown.

Servers are up revenue wise. Workstations isn't necessarily a major component of the rise of the joint category.

Market share would go up with all this? Don't think so. But the sales would skyrocket.

Skyrocket? Errrr. Stop selling 2+ year hardware and come out with something with current hardware... Uh? Yes. The Skyrocket is relative to the craptastic sales rate have repressed the Mac Pro into over time. In Dec 2013, sales skyrocketed in the EU after Apple could actually sell a Mac Pro due to their "most of 2013" product being way tooo old.

The is as much bozo product management as 'brilliant' product management.



[quote[ For people who actually earn money from work on computer, and want to use Apple ecosystem, or applications. For people who don't want to tinker with their computer a lot, and just use it. [/quote]

Plus I think Apple is getting the benefit of assuming the new role of the old "Nobody ever got fired for buying IBM". IBM was a big, solid, financially stable company. There is a "safer to buy from" notion that Dell, HP, Lenovo can't match. Apple sitting on top of $100B in cash so they probably won't implode in the next several years. HP ... are they even going to be called HP anymore? Lenovo appears to be China growth market one-trick pony that may be in trouble. Dell ... had to go private. Compared to that group Apple (as a company) looks rock solid stable.

Even IBM is selling Apple as a "safe choice".

And I don't think Apple cares about market share if they see in the end boost in sales of their products.

Apple does care about market shares of certain subsets. They don't care about gross overall aggregate market shares. For the subset of markets/products they target they do want to be a major player. (not necessarily the biggest, but major enough to get the major discounts on components for those products) But yeah it is handy to have sub 10% of PC market and be able to "hand back" portions you don't want anymore back into the 90% and crave out relatively small parts of the 90% as replacements.

P.S. There is a word from game developer that Fury Nano might be faster than Fury, and in the range of Fury X.

It doesn't say in the range of Fury X, just faster than Fury. In that broad sense the Fury is in the range of Fury X too if set the "in the range" bars to the right width. Fury X isn't radically faster than Fury; especially for the Fury versions that overclock.

".. Although looking at the bare specifications of the two cards would suggest they’d be fairly far apart in performance, this is not what we have found. Between 4K and 1440p the R9 Fury’s performance deficit is only 7-8%, noticeably less than what we’d expect given the number of disabled CUs. ... "
http://www.anandtech.com/show/9421/the-amd-radeon-r9-fury-review-feat-sapphire-asus/19

Probably similar issue with Fury Nano. Clocked down, but probably not loosing a whole lot because to large extent same foundation and features. The HBM being "faster than can leverage" means hardware gaps get smoothed over because the data is present when have time to use it. Hardware stays highly utilized most of the time. And likely few if any are trying to do heavily optimizations for the relatively small hardware gap between the products.

 

[deconstruct60]

It also can have Water Cooling. Which IMO would be a killer feature, for me at least .

For a system generally the size of the Mac Pro water cooling is dubious. For the Mac Pro specifically it is a Rube Goldberg solution in search of a problem. Primarily, water cooling is used to move heat from one dense location to another closer to the overall system's edge. The Mac Pro is small enough there are no "distance transfer" problems to solve.

The metal thermal core is better at moving heat than any PC water system. Just about every PC water system out there uses water to drop the heat off at metal component next to the fan. ( there are big iron air evaporators but that's a different class of thermal problem. ). The Mac Pro's thermal core go almost right up to the fan. Where is the long distance to cover? There isn't.

The Mac Pro could use a more efficient fan assembly. Either a bigger diameter ( about 1" would do) or stop using such a high percentage of the available diameter for a Wifi card and antenna. Or maybe a bit of both ( 1" taller and wider so proportions the same and perhaps move the WiFI electronics from being on top thermal core. )


The flat plate at the top of the Mac Pro moves no air. Probably used so that goofy folks don't drop stuff inside. Functionality wise though it a place can have clean RF 'view' to the outside world ( not encased in metal).

The Mac Pro only needs to move more air in a just as quiet fashion to be more thermally effective. That's it. No disco waterfall.

Classic PCIe cards need water because the "fan exit hole" is limited to the size of the narrow card edge. Water is needed to move the heat to a hole that actually round and aligned with what a fan can efficiently move air though. The Mac Pro doesn't even attempt to move heat out a card's narrow edge at all.

 

[Stacc]

I am waiting for the liquid nitrogen Mac Pro! Just imagine how many CPUs and video cards we could stuff in the tube then!

 

[koyoot]

My god, dec60, your posts are great lecture

I have to agree with you on the water cooling. I did not thought that best performance of water cooling we would get only from getting out the water of the case, or where it can be dissipated easily.

 

[SoyCapitanSoyCapitan]

Radeon Fury Nano game benchmarks are starting to appear. It looks about as fast as the 390X. A little slower than GTX980, much slower than the Fury X, 980Ti, or Titan X.

 

[goMac]

In alot of cases were Metal is called? Or Metal isn't being called very much by legacy (and released in last year or so ) Mac applications?




All Apple needs to do is hook foundation library services into it to make it run.

for the set of apps that minimize Apple libraries and compose their own "direct to OpenGL" graphics library .... then yeah.. If don't call Metal it won't show an improvement. But that is going to be a certian subset of applications. ( granted probably higher than normal usage by many Mac Pro users, but a subset none-the-less. )

There are issues with fully native Metal with some types of applications. It's not going to be what it's hyped up to be for a lot of apps. Especially in 3D applications.

 

[ActionableMango]

It also can have Water Cooling. Which IMO would be a killer feature, for me at least .

Apple tried that with the G5 PowerMac. It didn't go so well.

 

[koyoot]

Radeon Fury Nano game benchmarks are starting to appear. It looks about as fast as the 390X. A little slower than GTX980, much slower than the Fury X, 980Ti, or Titan X.

Where do you see them? I only see what has AMD provided.