Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
I'm so disappointed in the new mac pro mini

You may be better of getting a PC

I have two mac pros and will be keeping them until I will not be able to use them and than look at dell

Yes this sucks but apple is making a huge mistake and they will loose allots of photographers and videographers

I have 4 drives in both my macs and I hate external drives as they are much slower, and new drives cost a fortune

Sorry apple, but you lost a great supporter

If it is simply mechanical storage, you can find cheaper USB 3.0 or eSATA enclosures that will actually be faster than what you currently have. A few hundred dollars shouldn't be the difference between switching platforms completely.
 
Graphics cards.

Which ones? Are you telling me that every PCIe 3.0 card out there pumps data at the full 32 GB/s? I do not believe that, if that is what you are saying.

The new Mac Pro and a Thunderbolt drives/RAID setup (along with any legacy software a person owns) will be enough for many new buyers. For those that use additional PCIe cards, they can use Magma or Sonnet expansion chassis that will still be able to use their cards, and those chassis will be in the same racks that they were in when they were using their older Mac Pros.

A dual FirePro setup far outstrips any card currently in the now-old Mac Pros. The speed of the SSD is far greater than most current SSDs in the market. For whom they are intended, the FirePros w/OpenGL and OpenCL support will more than meet the needs of most professional users.

Apple is attempting to change the paradigm of the "trucks" that we use. Those that do not like it can buy HP or Dell or Boxx workstations or build hackintoshes.

I would not be surprised to see the general specs "bump up" as we move into the fall and holiday seasons and Intel brings other stuff to market.
 
**** you Apple for Destroying Mac PRO.

1. what if I want 4 SLI Quadro Workstation ?

2. what if I HATE AMD ?

3. you made it External and did not make full PCIE 16X 3.0 External PORT ? who CARES about Thunderbolt we HAVE EXTERNAL PCIE 16X 3.0 with Cables.

4. BLACK ? FOR REAL ? Apple Never Used black . I hate Black

5. Unified cooling ? WHO CARES we are not at Fashion show here !

6. Plastic ? REALLY ?

7. No Hotswapp Raid bay ? REALLY ?

8. No expansion slots ?

9. did you kill Steve Jobs to allow this **** ?

Maybe people should take a look at the MacRumors article Steve Jobs leaves Apple with four years of products. Its classic Jobs: the computer as appliance with little or no user serviceable or user-replaceable parts inside. Like a toaster its something you turn on and use, and if it breaks you send it to a service centre or get another one.

Now this has its good points and its bad points. If its something basic like a toaster and its cheap to make variants such as two and four-slice models then there are no obvious bad points. If its an item with a high cost of entry or replacement then closer attention needs to be paid to the bad points.

Having started over twenty-five years ago with an Apple IIe, then a Mac SE and followed in succession by a SE/30, a Quadra 800, a Power Mac 8500, a Power Mac G5, and most recently a 2008 Mac Pro, the bad points IMO are:

1) lack of internal drive bays. Having switched to SSD for the primary boot, applications, and user drive ~18 months ago and had two die in that time, I'd have expected a minimum of two internal 2.5" bays or PCIe flash-drive ports to allow for drive mirroring if desired. Two internal 3.5" user-accessible drive bays would have been even better -- particularly if there was the option to mount either 2 x 2.5" or 1 x 3.5" drive in each bay -- and for the naysayers arguing for thunderbolt, me and my partner have always hated the number of cables, power-cords, and the like emanating from my external drives.

2) not having twice as many DIMM slots as memory channels to allow some future expansion without having to throw away existing memory.

3) no obvious option for either a single inbuilt graphics card/chip or a user-replaceable graphics card. For reference: my all-time favourite Macs were the Quadra 800 and Power Mac 8500 models as they had (for the time) a reasonable onboard GPU and PCI slots for a gruntier graphics system as and when desired. With the new Mac Pro, it appears any options will be BTO up front with no possibility of updating/upgrading later unless buying a whole new system.

4) no obvious security slot for a kensington lock or similar. Granted this is more a deterrent for an opportunistic burglar than an absolute can't steal it setup, but unlike a toaster, a computer is an expensive item to buy or replace.
 
Last edited:
I thought, that the need for progress and innovation is part of an Apple user's character, but after 5 years (2 active and 3 reading) on Macrumors I got the sense, that a great number of Apple user just don't differ that much to other brand users regarding change.

"I will stick with Snow Leopard..." was something I watched with amusement.
I updated to Lion 14 days ago, and once you get used to the changes you see the progress. It's magnificent.
Just open your eyes and most important your mind.
 
I'm an editor using all the major nle's, also heavily into after effects and 3d packages including 3dsmax, maya, c4d and Zbrush. I also own smoke for compositing and have just begun (trying) to use davinci resolve on my sup'd to the max late 2011 MacBook Pro.

I've been waiting anxiously for the new Mac Pro as my current mbp doesn't cut the mustard.

Personally it makes almost perfect sense to me. I was a bit worried about the lack of nvidia but some posts on this thread and other research has led me to believe open cl will catch up.

I would read between the lines that apple approached nvidia with very testing terms with regards to buy in to their new desktop. I'm guessing that ati were more willing to take the hit as it would be an opportunity to get back in to a large portion of the pro market. It makes sense that apple wouldn't want to price themselves out if the market and would be looking for an economic partner route.

The processing power is great, it's good enough IMO for 3d work. It'll handle test renders and simulations well, the rest can go to render farms.

Black magic and the foundry are in which speaks volumes.

The expandability is very flexible and allows me to get top speeds or/and tons of storage attached, I can do it how I like.

Security is a great bonus given i can move it somewhere safe when I go on holiday. It's also a possible game changer for the industry given there may not be a cheaper version for studios that would have allowed them to replace all their workstations. Now no doubt they will opt for mainly iMacs with a few of these as the top machines. The game changer possibility is that I will be able to chuck mine in my rucksack and charge a little extra if they need the horse power. This redistributes the investment model for studios and freelancers and makes a lot of sense for all involved simply because there is less waste.

Will this last? I think this may depend upon apples ability to gain valuable supplier partnerships as and when new technology comes into play over the next 10 years.

I would expect this to cost around £5k or $7k
 
Regarding expandability, say this cylinder is purchased with a 256GB SSD. 6 months later the 256GB is outgrown so a 512GB SSD upgrade is purchased and installed. On any other Xeon workstation, the user would end up with 768GB of SSD storage, yet on this cylinder, the user ends up with 512GB SSD storage and a useless 256GB blade with nowhere to go! What an incredibly short-sighted design flaw.

Not only that, with just one PCIe SSD/flash-drive slot the user would have to make a bootable backup of the 256GB card on an external TB or USB drive, down the system, swap the 256Gb for the 512, reboot, then do another bootable backup from the external drive to the new 512GB -- thereby taking twice as long to complete the upgrade compared to backing up from the 256 to 512 or as you correctly said: using the 512 as expansion for the existing 256GB card.
 
I really hope this comes with a 768gb ssd or larger. I have found 512gb slightly limiting what with having windows on there too.

If its just a 512gb I might well just install windows on it (since as far as I'm aware you can't boot windows from an external drive) and then boot Mac OS from an external drive. Of course the advantage of this is that I can move to another workstation without even having to take my new Mac Pro along for the ride!
 
Wrongo...
thunderbolt 2 is 20 GB/s
PCIe 3 is 15.75 GB/s

Sorry you're mixing Thunderbolt's 20 Gigabits per second with PCIe 3's 15.75 Gigabytes per second. You need to divide the TB bandwidth by 8 to get the Gigabytes per second equivalent.
 
I thought, that the need for progress and innovation is part of an Apple user's character, but after 5 years (2 active and 3 reading) on Macrumors I got the sense, that a great number of Apple user just don't differ that much to other brand users regarding change.

"I will stick with Snow Leopard..." was something I watched with amusement.
I updated to Lion 14 days ago, and once you get used to the changes you see the progress. It's magnificent.
Just open your eyes and most important your mind.

It's not as simple as that. I moved to 107 and 108 as soon as I could because on the whole I like new things. There is always a trade off though, (old software and peripherals not working and of course a different way of doing things).
To say that those who stayed on 106 is something to watch with amusement sounds kind of stupid.
They have their reasons. Some of it is absolute cost and some is the continual problems that sometimes seem to occur to them.

----------

The game changer possibility is that I will be able to chuck mine in my rucksack ................

I would expect this to cost around £5k or $7k

That right there is something I've been wondering about. Quite a few times over the years I've wanted to take my Pro somewhere and decided against it due to th esheer heft of the thing. That said, I'd rather it was larger and light with some expansion, maybe 15" by 8".
 

Attachments

  • PCIe-3.jpg
    PCIe-3.jpg
    28.8 KB · Views: 85
If Apple wanted the video cards to be upgradable, they would have used a conventional PCIe slot.

----------



This is Apple we're talking about, yes? They've designed a computer with a proprietary video card installed in a proprietary socket and cooled by a proprietary heatsink. It's not the first such design from them, and in no previous such design were any upgrades offered.

Totally agree with this.

----------

Adobe has released press confirming OpenCL support already in Adobe CC

...which means that if, like me, you prefer your work to be on your local machine where its under your control then you're stuck with an older version of CS or need to look for different software.

----------

You also need to reduce T-Bolt 2 to 16 Gbps - since PCIe 2.0 has 20% error-checking overhead.

I stand corrected. :)
 
IT Doesn't Read....

I thought Mac Pro users would understand the lastest tech and how to expand. I was wrong.

I am baffled myself.

NEW Thunderbolt = 10Gbs

OLD sata = 6Gbs


um...I refuse to list all the improvements as people need to develop there reading skills.


And for those people talking about plastic take another look at the pictures, READ the description. its black polished aluminum or a black finished metal.

Not just for looks it also calculates really fast. :D
 
I am baffled myself.

NEW Thunderbolt = 10Gbs

OLD sata = 6Gbs

Old T-Bolt = 8 Gbps * 2
New T-Bolt = 16 Gbps * 1

One sata = 6Gbps
Two sata = 12 Gbps
Three sata = 18 Gbps
Four sata = 24 Gbps

PCIe 3.0 x1 = 8 Gbps
PCIe 3.0 x4 = 32 Gbps
PCIe 3.0 x8 = 64 Gbps
PCIe 3.0 x16 = 128 Gbps

(all speeds uni-directional)

So, if you put more than 2 fast SATA SSD drives per T-Bolt you are bandwidth-starved.
 
Old T-Bolt = 8 Gbps * 2
New T-Bolt = 16 Gbps * 1

One sata = 6Gbps
Two sata = 12 Gbps
Three sata = 18 Gbps
Four sata = 24 Gbps

PCIe 3.0 x1 = 8 Gbps
PCIe 3.0 x4 = 32 Gbps
PCIe 3.0 x8 = 64 Gbps
PCIe 3.0 x16 = 128 Gbps

(all speeds uni-directional)

So, if you put more than 2 fast SATA SSD drives per T-Bolt you are bandwidth-starved.

Is there no overhead for PCIe 3.0? And if you want to make TB "real world", isn't it unlikely that you'll actually get 3000MB/s from 4 SSDs anyways? SATA doesn't hit 6 Gb/s realistically does it?

Finally, what are the latency differences?
 
Last edited:
Is there no overhead for PCIe 3.0?

The numbers are after overhead is subtracted (PCIe 3.0 has 2 bits of overhead per 128 bits - less than 1%).


And if you want to make TB "real world", isn't it unlikely that you'll actually get 3000MB/s from 4 SSDs anyways? SATA doesn't hit 6 Gb/s realistically does it?

Something's confused about this question, especially since T-Bolt 2.0 has a theoretical peak of 2000 MB/s. Where'd 3 GB/sec come from?

SATA SSDs are benchmarking (that is, delivering real world results) in the 550 MB/s and higher range. That's pretty close to 6 Gbps (especially when you consider that the SATA protocol and error checking overhead that is *added* to the PCIe error checking and protocol overhead).
 
The numbers are after overhead is subtracted (PCIe 3.0 has 2 bits of overhead per 128 bits - less than 1%).




Something's confused about this question, especially since T-Bolt 2.0 has a theoretical peak of 2000 MB/s. Where'd 3 GB/sec come from?

SATA SSDs are benchmarking (that is, delivering real world results) in the 550 MB/s and higher range. That's pretty close to 6 Gbps (especially when you consider that the SATA protocol and error checking overhead that is *added* to the PCIe error checking and protocol overhead).

3 GB/s comes from 4 drives at 6 Gb/s.

So TB 2.0 really only 16 Gb/s of useable bandwidth? Where does the 20 Gb/s figure come from. Also, how can TB still be PCIe 2.0 if modern chips like Haswell are natively version 3.0? Are 4 lanes of 3.0 dedicated to each controller, or how does that work?
 
Any takers here?

Will this Mac Pro resolve the issue (Intel chipset actually) of true 23.x frames per sec playback? Will this Mac Pro (OSX issue actually) be able to play back HD audio streams found within media files? To date, the only way to get HD audio streams output is to load a non OSX operating system on a given Mac machine.
 
3 GB/s comes from 4 drives at 6 Gb/s.

Applying standard SCSI/SAS/SATA/FC 8b:10b encoding, 6Gbps = 600MB/s max theoretical bandwidth, so 4 x 6Gbps drives = 2400MB/s = 2.4GB/s

...and with benchmark results of 550MB/s each, it becomes 2200MB/s = 2.2GB/s actual.


So TB 2.0 really only 16 Gb/s of useable bandwidth? Where does the 20 Gb/s figure come from. Also, how can TB still be PCIe 2.0 if modern chips like Haswell are natively version 3.0? Are 4 lanes of 3.0 dedicated to each controller, or how does that work?

20Gb/s TB is the raw bandwidth before encoding and protocol overheads are applied.

-----
Edit:
Just because a bus is PCIe 3.0 doesn't mean its not backwards compatible to 2.0 or even 1.0. The fact that TB is PCIe 2.0 just means it will either auto-sense or use bandwidth on the PCIe 3.0 bus up to the limits of PCIe 2.0 and no higher -- and all downstream TB devices will be correspondingly limited to the max bandwidth of the TB bridge (PCIe 2.0).
 
Last edited:
Applying standard SCSI/SAS/SATA/FC 8b:10b encoding, 6Gbps = 600MB/s max theoretical bandwidth, so 4 x 6Gbps drives = 2400MB/s = 2.4GB/s

...and with benchmark results of 550MB/s each, it becomes 2200MB/s = 2.2GB/s actual.


Thus, since peak TB 2.0 bandwidth is 2.0 GB/s, you're not slowing down hardly any even with 4 drives. Also, 4x 550 MB/s SSDs in an array don't scal perfectly do they? That is, even if they are in RAID 0, you can't just add the individual performance numbers for peak performance of the array in RAID 0 can you?

All to say, it seems a little unlikely that anything above two SATA devices would be able to completely saturate the TB link.



20Gb/s TB is the raw bandwidth before encoding and protocol overheads are applied.

-----
Edit:
Just because a bus is PCIe 3.0 doesn't mean its not backwards compatible to 2.0 or even 1.0. The fact that TB is PCIe 2.0 just means it will either auto-sense or use bandwidth on the PCIe 3.0 bus up to the limits of PCIe 2.0 and no higher -- and all downstream TB devices will be correspondingly limited to the max bandwidth of the TB bridge (PCIe 2.0).

But if TB is 4 lanes of PCIe 2.0, where does the 20 Gb/s come from? 4 lanes of 2.0 is only 16 Gb/s anyways. Or are those numbers already including the overhead?

Also, why is TB limited to PCIe 2.0?
 
Thus, since peak TB 2.0 bandwidth is 2.0 GB/s, you're not slowing down hardly any even with 4 drives.

If you define "hardly any" as at least a 10% performance loss, true.


But if TB is 4 lanes of PCIe 2.0, where does the 20 Gb/s come from? 4 lanes of 2.0 is only 16 Gb/s anyways. Or are those numbers already including the overhead?

Marketing numbers seldom include actual real world performance - so the 20 Gbps raw number is marketed, and nobody mentions that the payload cannot be more than 16 Gbps.


Also, why is TB limited to PCIe 2.0?

T-Bolt 1 was PCIe 1.0 - because it wasn't any faster than PCIe 1.0 (in a 4 lane to 4 lane bridge).

We have very little info on T-Bolt 2, but its speed does match with a PCIe 2.0 4 lane to 4 lane bridge.

In theory, T-Bolt 2.0 could provide a PCIe 3.0 x1 or x2 interface. Or, it could provide a PCIe 3.0 x4 with pauses between packets - but there would be little actual advantage to doing that other than possibly being able to run at 20 Gbps instead of 16 Gbps depending on the encoding.

In truth, though, T-Bolt is an opaque, proprietary interface that is very difficult to get any facts on unless you are a licensee. We don't know if the host side is a PCIe 2.0 or 3.0 part, nor if the device side presents a PCIe 1.0, 2.0 or 3.0 bus.
 
Last edited:
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.