This is not a "Pro" machine.

[flat five]

FWIW, I would like to see it because it is an interesting idea from the perspective of any easy-DIY cluster for modular performance growth...of course, we could also see Apple not wanting that because it slows/reduces hardware upgrade cycles.

that would be a bummer if they didn't allow it strictly for that reason.. and they probably won't because you can just as easily (or easier since there's not cable involved) connect wirelessly (as in you follow the same exact steps to hook up wireless or via firewire).. i guess in many situations, you would take a major performance hit but with stuff like rendering, you wouldn't.

i hope it's more along the lines of it being an additional 'feature' they can announce when 10.9 comes out..

 

[Tesselator]

that's as a target disk tho.
the capability in question is IP over thunderbolt.

Which is why I said someone needs to write protocol.

Yes, this has been the underlying point on the "loss of internal bays" discussion splinters: this level of performance capability that's been baked into the Mac Pros for years, and configrations can take it up by another 50% without major grief. The "least bad" option today appears to be to use USB3...something that really should have been OEM on the 2912 via a PCIe card, but that's a tangential gripe of Apple's chronic neglect of the MP platform.

I don't follow this at all. And from what I do understand of it I don't agree. We had four SATA2 connections good for 300MB/s and not daisychainable. Now we have USB3 good for 500MB/s and which ARE daisychainable. Rotational devices max out at around 200MB/s so generally speaking we can put three drives (in RAID0) on each port for relatively the same performance as three drives in the old SATA2 (MP5,1) ports. But we have four of the USB3 ports so that's 12 total drives available and all you gotta do is add a $125 enclosure per.

Should'uv, would'uv, could'uv, if that's really the "tangential gripe" then that's pretty retarded.

Depends on what you consider to be 'networking': TB can be used in target disk mode, and TB can be used with an Ethernet or Fibre Channel adapter too.

....but the idea of hooking up two autonomously booted/managed Macs with just a TB cable and having it work like Ethernet? Not so per Apple marketing, nor even any well-informed discussions on the underlying technology.

Yeah there needs to be some TP/IP protocols added in (which I don't think exist yet) before that can work. Once in place it should work OK tho. (guessing)

FWIW, I would like to see it because it is an interesting idea from the perspective of any easy-DIY cluster for modular performance growth...of course, we could also see Apple not wanting that because it slows/reduces hardware upgrade cycles.

Ya, with nic binding we can get 200MB/s from the built-in ethernet ports or 100MB/s without binding but the 1GB/s or 2GB/s that TB1 and TB2 offer respectively looks too good to pass off.

 

[-hh]

Which is why I said someone needs to write protocol.

Which might be another bottleneck, or incur some other issue.

I don't follow this at all. And from what I do understand of it I don't agree. We had four SATA2 connections good for 300MB/s and not daisychainable.

Five if you remember the empty optical bay, and six if you're willing to pull the included optical.

Now we have USB3 good for 500MB/s and which ARE daisychainable. Rotational devices max out at around 200MB/s so generally speaking we can put three drives (in RAID0) on each port for relatively the same performance as three drives in the old SATA2 (MP5,1) ports. But we have four of the USB3 ports so that's 12 total drives available and all you gotta do is add a $125 enclosure per.

But since the same can be done with a MP 5,1 with an eSATA or USB3 PCIe card, this isn't a meaningful improvement in capability.

Plus there's only four USB3's if one has no other USB peripherals to consider...keyboard, mouse, tablet, local printer, scanner, film scanner, thumb drive, iPhone, flash memory reader, drive toaster, etc, etc. Sure, some of these can now go wireless, and we can throw a hub on too, but it realistically means that only one or two external drive stacks are going to exist in the wild...and from a Pareto Principle, that's more likely anyway and from that perspective, the most common use case is going to be just one ... and with the 5,1 the first stack was figuratively "free" because of the internal capability.

Should'uv, would'uv, could'uv, if that's really the "tangential gripe" then that's pretty retarded.

Actually, it is just a comment on the canary in the coal mine, as this was YA validation showing that the MP has been at the very bottom of Apple's priority list for hardware maintenance. YMMV if the changes between the 5,1 and 6,1 is going to change this for the better or not...I believe not.

Ya, with nic binding we can get 200MB/s from the built-in ethernet ports or 100MB/s without binding but the 1GB/s or 2GB/s that TB1 and TB2 offer respectively looks too good to pass off.

Technologically, sure ... but the basic capability is already in the marketplace in the form of a Fibre Channel adapter card. As such, Apple's business centric decision making processes of late are too likely IMO to ignore the idea of baking it in for free, even though it would be a product differentiator.


-hh

 

[Tesselator]

Which...

Yeah, all fairly valid points. I don't agree with some of them but most are just a matter of balance and bias - and perhaps some product awareness. When one stops to consider all things together the MP6,1 can do everything the MP5,1 can do but a few expansions might need a TB2 PCIe card adapter which are a little pricey right now. In the end with PCIe expansion boxes on both systems the MP6,1 actually comes out quite a bit cheaper if you don't count the base system price - and configurably more dynamic as well.

I guess this gap will widen over time too as PCIe expansions from TB are just starting to hit the market and the expansion systems for slotted system have been around for a long time and not likely to come down in price. Competition will likely contribute to this as well with 6 companies already in the running plus a score more on the horizon readying up.

• http://www.magma.com/expressbox-3t
• http://www.sonnettech.com/product/echoexpresschassis.html
• http://eshop.macsales.com/item/Other World Computing/HELIOSTB1/
• http://www.netstor.com.tw/_03/03_02.php?MTEy
• http://www.maxexpansion.com/kits/thunderbolt-host-target
• http://www.mlogic.com/products/mlink

Hopefully Apple will also solidify this by setting a low entry price for the new machines. Considering that not even 10% of MP users need or want even half of the expansion we've been talking about till now it should be a pretty popular system. We'll see I guess. It'll be interesting to watch how the new MP gets used and read the comments as people purchase and incorporate the system into their work environments.

Which might be another bottleneck, or incur some other issue.

Here's a 10Gb networking TB solution. Looks a little pricey but hey... so was 4GB FC too...
http://www.mlogic.com/products/mlink-10gige-solution

 

[-hh]

Yeah, all fairly valid points. I don't agree with some of them but most are just a matter of balance and bias - and perhaps some product awareness.

The problem with trying to figure out the balance are the self-contradictions in the discussion. For example, the EDIT: SIX (6) TB ports are being touted as the 'better' yet we are then told this:

Considering that not even 10% of MP users need or want even half of the expansion we've been talking about...

So if it really is true that the MP demographic only extremely rarely invokes any capability growth, then why did Apple spend the money to put six (6) TB ports on the 6,1? From a business case perspective, it costs more to do, and even the remaining 10% would have been saturated with two.

We can't claim (with a straight face) that it is for displays, since TB daisy-chains and video's on a separate data channel.

What's plain to me is that I'm being lied to. I simply haven't been able to prove who the liar is.

Here's a 10Gb networking TB solution. Looks a little pricey but hey... so was 4GB FC too...
http://www.mlogic.com/products/mlink-10gige-solution

Yes, as I explicitly said a few weeks ago, but which I'll now re-emphasize: if the 6,1 was really intended to be a network-centric workstation, its specifications should have nothing less than 10GB Ethernet port(s) on it. Of course, the problem with such an approach is that it would compete with TB.


-hh

 

[subsonix]

Yes, as I explicitly said a few weeks ago, but which I'll now re-emphasize: if the 6,1 was really intended to be a network-centric workstation, its specifications should have nothing less than 10GB Ethernet port(s) on it. Of course, the problem with such an approach is that it would compete with TB.

10GbE is typically an add-on regardless of what platform you use, sfp+ is not used by regular internet routers. It makes sense on a LAN if you buy into the infrastructure.

 

[Tesselator]

Yeah, all fairly valid points. I don't agree with some of them but most are just a matter of balance and bias - and perhaps some product awareness.

The problem with trying to figure out the balance are the self-contradictions in the discussion. For example, the four (4) TB ports are being touted as the 'better' yet we are then told this:

Considering that not even 10% of MP users need or want even half of the expansion we've been talking about...

So if it really is true that the MP demographic only extremely rarely invokes any capability growth, then why did Apple spend the money to put four (4) TB ports on the 6,1? From a business case perspective, it costs more to do, and even the remaining 10% would have been saturated with two.

We can't claim (with a straight face) that it is for displays, since TB daisy-chains and video's on a separate data channel.

What's plain to me is that I'm being lied to. I simply haven't been able to prove who the liar is.

I'm not sure I see any discrepancy or reason for confusion there. The "balance and bias - and perhaps some product awareness" have nothing to do with the discussion. It's how a potential user balances their budget and configuration, bias toward storage, networking speed, or etc, and what products they're aware of when they plan the config.

Also there are six (6) TB ports, not four (4), plus an HDMI port. Heck, for monitors there are even USB3 to HDMI adapters these days - and they're fast enough for most graphics work. I was using one on my Dell XPS-12 laptop running PS-CS6 and LW3D the other day and it seemed the same as regular HDMI or DVI to me. I didn't try any games or anything tho.

I guess I'm not understanding why the (probable) fact that only about 10% of people upgrade or expand their machines very much is in conflict with anything said anywhere here or by Apple. Machines always have a certain expansion potential and the MP6,1 is no different that I can see. And it's just that... a potential.

Is TB2 "better"? Don't we have to define better-that-what in order to ask that question? As the MP6,1 is configured from Apple and as shown at WWDC there are already 2 strong GPUs installed so almost no one needs to ask if it's better than PCIe x16 v3 right? What's left? Storage, networking, something else? Well, that's why I said the bit about balance, bias, and product awareness. Everytime I do a search and crunch a few numbers TB2 comes out ahead or the same as what MP5,1 had - although more expensive in some cases. And, in the cases where it was both more expensive and a commonly employed expansion (such as large storage) there are cheaper ways of implementing it - like via USB3.

What is it you think you're being lied to about?

 

[blackhand1001]

I'm not sure I see any discrepancy or reason for confusion there. The "balance and bias - and perhaps some product awareness" have nothing to do with the discussion. It's how a potential user balances their budget and configuration, bias toward storage, networking speed, or etc, and what products they're aware of when they plan the config.

Also there are six (6) TB ports, not four (4), plus an HDMI port. Heck, for monitors there are even USB3 to HDMI adapters these days - and they're fast enough for most graphics work. I was using one on my Dell XPS-12 laptop running PS-CS6 and LW3D the other day and it seemed the same as regular HDMI or DVI to me. I didn't try any games or anything tho.

I guess I'm not understanding why the (probable) fact that only about 10% of people upgrade or expand their machines very much is in conflict with anything said anywhere here or by Apple. Machines always have a certain expansion potential and the MP6,1 is no different that I can see.

Is TB2 "better"? Don't we have to define better-that-what in order to ask that question? As the MP6,1 is configured from Apple and as shown at WWDC there are already 2 strong GPUs installed so almost no one needs to ask if it's better than PCIe x16 v3 right? What's left? Storage, networking, something else? Well, that's why I said the bit about balance, bias, and product awareness. Everytime I do a search and crunch a few numbers TB2 comes out ahead or the same as what MP5,1 had - although more expensive in some cases. And, in the cases where it was both more expensive and a commonly employed expansion (such as large storage) there are cheaper ways of implementing it - like via USB3.

What is it you think you're being lied to about?

When it comes to professional workstations its far more than 10% are eventually upgraded or expanded. I'd say its close to 100% after a few years.

 

[Tesselator]

When it comes to professional workstations its far more than 10% are eventually upgraded or expanded. I'd say its close to 100% after a few years.

Yes, probably. But the context was this:

Hopefully Apple will also solidify this by setting a low entry price for the new machines. Considering that not even 10% of MP users need or want even half of the expansion we've been talking about till now it should be a pretty popular system. We'll see I guess. It'll be interesting to watch how the new MP gets used and read the comments as people purchase and incorporate the system into their work environments.

Where two things are implied there. One, that whatever expansion we're talking about is part of the initial configuration and two, that we were talking about total expansion potential like 12 or 18 GPUs, 12 Rotational HDDs, 6-drive SSD RAID0, 10G Networking, Audio interfaces, and so on. So I ask you then; What percentage of MacPro users will "need or want even half that" initially? I guess it's less than 10% myself. No?

 

[-hh]

When it comes to professional workstations its far more than 10% are eventually upgraded or expanded. I'd say its close to 100% after a few years.

Agreed. Where I suspect that the "Only 10%" statistic is probably coming from is the 80% of Macs that are laptops, which is not germane to a discussion on workhorse desktops.

If there was a valid scientific survey done specifically on Mac Pros, I'd like to see the published reference...


-hh

 

[uncleMonty]

Good news!

I just got an address and a date for Tuesday. I plan to bring a camera to record the result, because as far as I know, this will be the first documentation of Thunderbolt moving more than 1300MB/sec... and it sounds like this will do more than 2000MB/sec, so it will be a world premiere!

I'm seriously quite excited for this.

So hey whatever happened with this, I know you missed a flight and had to wait a week, but did you end up visiting Transeau at his workplace in LA and confirming his TB throughput?

OP will surely deliver!

 

[wonderspark]

So hey whatever happened with this, I know you missed a flight and had to wait a week, but did you end up visiting Transeau at his workplace in LA and confirming his TB throughput?

OP will surely deliver!

He got in trouble, and the meet was cancelled due to a complete lockdown at the facility. I've not mentioned anything, just to keep the (hopefully) settling dust from being disturbed.

I got the impression that I will not be allowed into the facility, or at least not into the lab where this equipment is located.

Apparently, he was caught running a speed test on said equipment, and was seen doing so by someone who reported it, and he was called onto the carpet to explain himself.

 

[uncleMonty]

He got in trouble, and the meet was cancelled

Thanks for the explanation. That's too bad Transeau got in trouble - what kind of jackass reports a coworker to the boss for running a speed test?!! I hope the trouble passes.

 

[wonderspark]

It sounded like the trouble was temporary, and things will return to normal eventually. There is more to say, but I think it's best that he talk about it when he is able.

 

[freejazz-man]

ironic that people complain that the new mac pro looks like a nyc subway trashcan considering that if it actually was you'd be able to fit something like 1000 SATA hard drives in there. No more taking out the optical bay guys!!

 

[reorx]

Some people just want to post cute puppy pix on Facebook and uplaod an occasional video to Youtube. Perfect machine for it. At a reasonable price, they'll sell like hotcakes to the Semi-ProSumer crowd.

My MacPro 3,1 does a fine job at browsing cat pics on reddit and its still better than my workstation at my job as a developer at a Fortune 200. Every situation a "professional" finds themselves in is different, and to say that they aren't is just silly. At work, nothing but executive software goes on local storage - all workspaces are on the network: better monitoring, manageability and uptime...

Do we upgrade video cards? No
Do we upgrade local storage? No
Do we upgrade CPUs? No

Are we professionals that use computers and software to make money? Yes.

 

[MacVidCards]

My MacPro 3,1 does a fine job at browsing cat pics on reddit and its still better than my workstation at my job as a developer at a Fortune 200. Every situation a "professional" finds themselves in is different, and to say that they aren't is just silly. At work, nothing but executive software goes on local storage - all workspaces are on the network: better monitoring, manageability and uptime...

Do we upgrade video cards? No
Do we upgrade local storage? No
Do we upgrade CPUs? No

Are we professionals that use computers and software to make money? Yes.

Back in my day, we could install up to 6 (yes, SIX) SATA drives without spending an additional nickel paying some company Apple made a deal with to sell TB controllers for.

Yep, that's right, we could choose to install 1 or 6 drives or anywhere in between, and just pay for the drives themselves. No external enclosures, sleep issues, TB cables or power supplies. Sounds like a fairy tale now but that's how it used to be.

 

[flat five]

Back in my day, we could install up to 6 (yes, SIX) SATA drives without spending an additional nickel paying some company Apple made a deal with to sell TB controllers for.

Yep, that's right, we could choose to install 1 or 6 drives or anywhere in between, and just pay for the drives themselves. No external enclosures, sleep issues, TB cables or power supplies. Sounds like a fairy tale now but that's how it used to be.

you're out of your mind if you think you weren't paying for the the drive bays and connectors and large enclosure. i'd guess it's even more $ than a thunderbolt enclosure and cable will eventually cost..

 

[tuxon86]

The mac pro can acommodate 6 drives internally. No need for large enclosure, bays or connector beside $2.00 sata cable.

You don't really own a Mac Pro now do you?

 

[freejazz-man]

I've never seen someone as pissed by something so trivial than macvidcards. I think you are going to have to get over it, dude!

I really own a mac pro and I don't have 6 disks in there! shoooocking

 

[slughead]

Back in my day, we could install up to 6 (yes, SIX) SATA drives without spending an additional nickel paying some company Apple made a deal with to sell TB controllers for.

you're out of your mind if you think you weren't paying for the the drive bays and connectors and large enclosure. i'd guess it's even more $ than a thunderbolt enclosure and cable will eventually cost..

Wait, you're saying the MARGINAL COST of a larger computer case and onboard sata controller (Vs a smaller case and no SATA controller on the mobo) is going to be MORE than adding

1) Thunderbolt Controller
2) Thunderbolt cable
3) Thunderbolt -> SATA controller
4) External enclosure
5) External PSU

I think you're supposed to drink the kool-aid, not snort it. Do you know what kind of ECONOMY OF SCALE the existing Mac Pro has using a near-standard Intel motherboard? This is something Apple will no longer enjoy (unless you think that the nMP sales will rival all other Xeon motherboards combined). It's probably such that removing the SATA controller would cost more to produce than keeping it there!

Yes, the new mac pro uses a crappy plastic case (see edit). Is it going to be cheaper than the high-quality aluminum of the old case? Yes. Is it going to be THAT MUCH cheaper? No. The move to a more proprietary motherboard may even have the nMP costing more to produce than the existing one (ignoring the price of the video cards).

EDIT: I was wrong, as others have stated the nMP uses aluminum. This means the cost difference between this case and the old one will be even LESS.

 

[slughead]

My MacPro 3,1 does a fine job at browsing cat pics on reddit and its still better than my workstation at my job as a developer at a Fortune 200. Every situation a "professional" finds themselves in is different, and to say that they aren't is just silly. At work, nothing but executive software goes on local storage - all workspaces are on the network: better monitoring, manageability and uptime...

Do we upgrade video cards? No
Do we upgrade local storage? No
Do we upgrade CPUs? No

Are we professionals that use computers and software to make money? Yes.

I'm glad you can make do with what amounts to a low-power terminal at your job. Can you at least acknowledge that for many (most?) Mac Pro users, the lack of internal storage will create an unnecessary expense?

 

[tuxon86]

I've never seen someone as pissed by something so trivial than macvidcards. I think you are going to have to get over it, dude!

I really own a mac pro and I don't have 6 disks in there! shoooocking

No, but yo could if you needed to at no extra cost except for the drives themselves and a couple of $2.00 sata cables. That's his point.

 

[GermanyChris]

I'm glad you can make do with what amounts to a low-power terminal at your job. Can you at least acknowledge that for many (most?) Mac Pro users, the lack of internal storage will create an unnecessary expense?

This is the Corporate v Jobber/Independent debate.

In big corps there will be little not issue integrating these in the infrastructure for independents it's different.

 

[mcnallym]

Yes, the new mac pro uses a crappy plastic case. Is it going to be cheaper than the high-quality aluminum of the old case? .

New Mac Pro Case is machined aluminum not Plastic. Why do some people insist on calling it Plastic when one of the few facts out there is that the case is machined aluminum.

No, but yo could if you needed to at no extra cost except for the drives themselves and a couple of $2.00 sata cables. That's his point.

Why do you need Sata Cables to install 6 drives in a Mac Pro.

The 4 sleds if you connect the drives into them slot straight onto Power/Sata connectors on the main board, so why would you need a sata cable.

If you want to take fill the two optical bay then again there are already cabling run through for 2 devices Power/Sata that you plug straight onto the Disk.

At least that is what is inside my Mac Pro 2010, don't know how they work in other peoples Mac Pro's

The only cables I have had to purchase was 2 GPU Power Cables for my GTX680.

If you can't get the basics right then people will question what else you post!