Don't get too bent out of shape man, how was I supposed to know which program you use?
Plus, I said something LIKE it. I was insinuating that a lot of people are going the GPU not CPU route.
Last edited:
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.
nMP pricing compared to 2012
- Thread starter VirtualRain
- Start date
- Sort by reaction score
Don't get too bent out of shape man, how was I supposed to know which program you use?
Plus, I said something LIKE it. I was insinuating that a lot of people are going the GPU not CPU route.
Its great that the USB3 enclosures are much cheaper. Then there is TBv1 vs. TBv2 option. TBv2 should drive down the TBv1 device costs.
I'm not sure if there are TB2 enclosures on the market yet, or just announcements but they do look like 4-figure investments from what little I have found with a price attached.
My issue with USB3 is of course speed. You are buying a nMP which is getting 1200 MB/s (down from the 1250 they originally said) from its internal pcie SSD. Again, why I wish you could put in a 2nd internal ssd.
A USB3 enclosure I think you'll get 100 MB/s with 7200HDD's, and 200 MB/s with SSD's. TB and USB3 are only close performance wise when you are only plugging in one drive. When you have 3 or 4 SSD's in the TB enclosure you will be getting 700-800 MB/s. You might get to 300 MB/s with multiple SSD's and USB3.
The J4 looks like a good option. It might get even cheaper when the Pegasus TB2 enclosures show up.
http://www.macworld.com/article/2039427/how-fast-is-usb-3-0-really-.html
I'm not sure if there are TB2 enclosures on the market yet, or just announcements but they do look like 4-figure investments from what little I have found with a price attached.
My issue with USB3 is of course speed. You are buying a nMP which is getting 1200 MB/s (down from the 1250 they originally said) from its internal pcie SSD. Again, why I wish you could put in a 2nd internal ssd.
A USB3 enclosure I think you'll get 100 MB/s with 7200HDD's, and 200 MB/s with SSD's. TB and USB3 are only close performance wise when you are only plugging in one drive. When you have 3 or 4 SSD's in the TB enclosure you will be getting 700-800 MB/s. You might get to 300 MB/s with multiple SSD's and USB3.
The J4 looks like a good option. It might get even cheaper when the Pegasus TB2 enclosures show up.
http://www.macworld.com/article/2039427/how-fast-is-usb-3-0-really-.html
Not necessarily because of TBv2. It is likely going to be common that TBv2 discontinue TBv1 products. There may be a temporary blimp as "last years" model is discounted to remove but long term discontinued products disappear; at least as new products. v1 is going to be far more a proxy for being used/old. And yes it is quite normal for used/old products to drop in price. (but again there aren't more of them being made so again this temporary although a much longer temporary. )
So far Intel is not making some radical change to TBv1 controller parts and there certainly isn't gobs of unused inventory floating around.
If start doing somewhat random/concurrent access to HDDs there is no speed problem.
Those spec drops are likely indicative of what it isn't happening. Unless, snarfing bandwidth away from something else there is no additional bandwidth to get. There don't particularly seem like there is extra power to go around either.
Actually a good illustration of why not to keep the TBv1 version around. The TBv2 version will likely drop around same price and this 'v1' version will disappear. Folks are quite likely to throw 2-4 SSDs at the J4. TBv2 is far more able to keep up with that.
I wish we could swap out a GPU for another SSD blade. I do heavy photo editing/compositing which requires CPU/RAM/Fast Drives. To buy even the base model I'm going to have to go TB/TB2 with an external RAID setup, which adds a minimum $1k to my cost, and upgrade the paltry 256GB SSD to the 512GB. So my $2,999 nMP is up to $4,500.
I've said before unless it's somehow 2x faster than the i7 BTO iMac there's no reason for me to jump to a nMP. I really don't like AIO designs but I'm also not going to blow $1k just to NOT have an iMac.
I do heavy photo editing as well. Having your working files on SSD makes a huge difference which is why this new Mac Pro is so appealing. I intend to max out the internal SSD to 1TB and put my archived photos/work on large 3-4TB drives (something I started doing a few years ago).
Also, using RAID for archival storage isn't wise, especially with larger 3-4TB drives.
Why not do what I'm planning?... run your day-to-day off SSD and archive stuff to large external drives (USB3). (when one fills up put it in a drawer and start filling another one).
I think it's an unnecessary requirement to have everything you've ever created "online" and accessible at any given moment. You're only adding unnecessary operating hours to that media you seldom access and consuming unnecessary power to keep it available on a second's notice. If I want a photo or project I worked on earlier this year, it's on my latest attached external. If I want one I took in 2010, it's on a drive in my drawer that I can hook up to my machine in 10 seconds. I don't need to keep that disk spinning for the last 2 years just to save a few seconds of time. Is this why you guys need such large RAID arrays for? If so, I think you're crazy!
I'm not THAT crazy
I upgraded my main MP drive to a Samsung 840PRO SSD (which is great even with my limited SATA 3G speeds) and I have my design/photo files on a 3TB RAID 0 internally, backed up to external matching systems daily.
I wish I could just move my SSD over and add as usual but I can't which is why I was stating the need for a ton of external drives (My PSD working files can be upwards near 1GB in size for large format jobs). I just wasn't planning on having two sets of drive arrays externally for my next mac pro.
I wouldn't mind using cheaper USB3 alternatives for my backup but definitely not for my main data drive which I need to be as fast/faster than what I'm currently working with internal RAID. So TB is my only real option which is $$$.
I upgraded my main MP drive to a Samsung 840PRO SSD (which is great even with my limited SATA 3G speeds) and I have my design/photo files on a 3TB RAID 0 internally, backed up to external matching systems daily.
I wish I could just move my SSD over and add as usual but I can't which is why I was stating the need for a ton of external drives (My PSD working files can be upwards near 1GB in size for large format jobs). I just wasn't planning on having two sets of drive arrays externally for my next mac pro.
I wouldn't mind using cheaper USB3 alternatives for my backup but definitely not for my main data drive which I need to be as fast/faster than what I'm currently working with internal RAID. So TB is my only real option which is $$$.
I do heavy photo editing as well. Having your working files on SSD makes a huge difference which is why this new Mac Pro is so appealing. I intend to max out the internal SSD to 1TB and put my archived photos/work on large 3-4TB drives (something I started doing a few years ago).
Also, using RAID for archival storage isn't wise, especially with larger 3-4TB drives.
Why not do what I'm planning?... run your day-to-day off SSD and archive stuff to large external drives (USB3). (when one fills up put it in a drawer and start filling another one).
I think it's an unnecessary requirement to have everything you've ever created "online" and accessible at any given moment. You're only adding unnecessary operating hours to that media you seldom access and consuming unnecessary power to keep it available on a second's notice. If I want a photo or project I worked on earlier this year, it's on my latest attached external. If I want one I took in 2010, it's on a drive in my drawer that I can hook up to my machine in 10 seconds. I don't need to keep that disk spinning for the last 2 years just to save a few seconds of time. Is this why you guys need such large RAID arrays for? If so, I think you're crazy!
I'm not THAT crazy
I upgraded my main MP drive to a Samsung 840PRO SSD (which is great even with my limited SATA 3G speeds) and I have my design/photo files on a 3TB RAID 0 internally, backed up to external matching systems daily.
I wish I could just move my SSD over and add as usual but I can't which is why I was stating the need for a ton of external drives (My PSD working files can be upwards near 1GB in size for large format jobs). I just wasn't planning on having two sets of drive arrays externally for my next mac pro.
I wouldn't mind using cheaper USB3 alternatives for my backup but definitely not for my main data drive which I need to be as fast/faster than what I'm currently working with internal RAID. So TB is my only real option which is $$$.
Doesn't a pair of spinning drives in RAID0 do about 250MB/s? Isn't that within the performance of USB3? Why TB?
IMHO, the only spinning disk setup that's worthy of TB is the WD Velociraptor Duo which can get close to SSD like STRs of 400MB/s.
Those expensive 4-5 bay TB enclosures make little sense to me unless you intend to load them up with SSDs and even then there's the relatively affordable Promise J4 for that. RAID0 of four big disks adds little reward and lots of risk, large RAID5 arrays are extremely risky if you ever need to rebuild. Maybe RAID10 makes sense, but I think its much more sensible to run a RAID0 array of dual drives in a USB3 enclosure and back it up occasionally (to another USB3 enclosure) than to run RAID10 in a TB enclosure. I think RAID for bulk storage is just a bad idea in general. People do it because they can, not because it's wise.
My view is that people are much better off using SSDs where speed is important and single large disks for bulk storage. And I just can't believe that one can't get by on 1TB of SSD space for any kind of tasks you might need to do on a weekly basis. If 1TB is not enough, you'd make more sense if you told me you needed to buy a Promise J4 and more SSDs than a TB enclosure for spinning disks.
Understandable, but what's necessary/unnecessary varies by use cases, and the way that Apple structured the data format for iPhoto, its employment of one gigantic library (database) once it was split into smaller libraries, it traditionally could not be "re-merged".
As such, if you ever wanted to be capable of tasks like a longitudinal reach into your stock, you were at least forced to keep one instance of a complete "Master" library...and if it was "too big" to work on a laptop, you were forced to make a fork, where any post processing work done in the fork could not be seamlessly brought back into the master.
True, there have become some management tools to work with multiple libraries, most noteably
Fat Cat Software's "iPhoto Library Manager" ... but until recently, these were "ONE WAY" in that they were OK to split big to small, but there was simply no good way to ever merge small library instances back together into a 'big' (common) database.
As such, a common approach became "stop overthinking it ... just buy a bigger HDD and leave it as one big damn library" for all of the headaches of document retention, revision control, data management, etc.
Is this why you guys need such large RAID arrays for? If so, I think you're crazy!
Perhaps the answer there is that there's finally a solution to the above discussed libraries management problem, because Aperture ($80) is now advertising a 'Library Merge' feature ... as is now also iPhoto Library Manager ($30) too.
If it works as seems to be being advertised, the implications are that libraries can be easily split -and- merged, which means that the old one-way workflow trap has been (finally) abolished.
I haven't fully researched if this does fully solve the concerns and limitations that I've discussed above, but it does look fairly promising ... and if it really is the viable solution, then this simply becomes a matter of getting the word out about it.
Of course, we can also provide a bit of a grumble because Apple didn't bother to include this 'merge' feature within iPhoto...it looks like it is being employed as an 'upsell' feature for Aperture ($80).
-hh
I'd expect that a 4-disk RAID0 over TB would be in the 450-500 MB/s range, and if one just uses 1TB 3.5" HDDs, you're talking about it costing $300 (before enclosure), rather than ~$1200 for 2TB worth of SSDs, so its 1/4th the cost and the trade-off is the reliability bugaboo...
...to address reliability question, one can just ignore it by saying "that's what backups are for", since SSDs aren't perfect either.
Or one could make a strategic investment by buying standby hardware. For the above example, buying four spare 1TB HDDs would afford a 100% redundancy while still providing a 50% cost savings vs SSDs.
Personally, I've only lost one RAID within the past five years so our system risk trades analysis priorities/rankings can be a YMMV.
Or they do it for the cost savings -vs- the risk, as per the above illustration.
-hh
I use Aperture and generally create a new Library every 6 months. This way, the active library I'm working in, can fit on my SSD. The historic libraries go on a large drive for archival and when I need it, (which is increasingly rare as it gets older) I plug it in (or I can leave the latest archive attached). I have yet to need to merge, but I think you're correct in stating that Aperture does provide that functionality.
Yeah, of course you can do this, but as I said above, I think most people do it because they can, not because it's the best solution. No one needs SSD like transfer rates to access the occasional archive. That's gold plating to the extreme and of course it's going to be costly… a completely unnecessary expense. So I don't have any sympathy for anyone that says they MUST have a Thunderbolt enclosure for accessing Terabytes of archived data files that they touch once in a blue moon. If you're regularly going to your spinning disks for your project files, then you need a bigger SSD… period!
I suspect if people we're more realistic with their true storage needs, and stopped letting how they did it in the past blind their vision, they would find that a new Mac Pro with maxed out SSD and even something as simple as one of these HDD docks, was all they really needed to be super productive…
EDIT: I should say, I'm not advocating one of these docks, but they help make my point. Of course there are a variety of choices for attractive USB3 enclosures that will both protect the drive and enable a quick connection, which is what I would use, but you get the point.
Last edited:
Uh....no...Single spinning disks rarely go over 120-180MB/s, and thats even pushing it unless you go into SAS drives which still is pushing it. 4-5 drives is needed for accetable throughput when editing even one or two compressed streams of high quality video (ie: ProRes HQ, REDCODE, ArriRAW, etc.). It doesn't seam like you've looked at the big picture.
Accessing terabytes of achived data maybe is weird but regularly accessing terabytes of data isn't weird and not wanting to spend large amounts of money on SSD's is understandable. Also RAID's can be better suited for certain tasks over SSD's obviously. I just don't think its fair to generalize anyones storage needs. My dad thinks its rediculous I have over >6TB's worth of external hard drive space but then again he can't tell me what a GPU is or how to edit a film.
Just gonna leave these screen caps here. I saved them since its hard to find old specs in a neat format as when they are new. Feel free to reference them. Pricing seams fair with the new Mac Pro although you really have to take into account the completely different upgradability factor. At least the new one comes with a decent amount of RAM, I think 12GB's is sufficient. GPU's definitely just 2012's options, and I don't even really like AMD.
Attachments
Last edited:
I dunno man… an hour of ProRes HQ is like 80GB. It's crazy to want to edit that on spinning disk media. SSDs are now cheaper than HDs were just 7 years ago (1TB for $500). There's really no excuse for anyone to be working off HDs as their primary storage medium anymore.
Having said that, HDs still have a place for archiving content or storing stuff that doesn't benefit from high speeds or frequent access… like your video and audio collection or last years photos.
RAID is another matter… with SSDs and very large fast high-density 3-4 TB drives, the benefits of RAID have really diminished. Even RAID0 doesn't really make sense except perhaps in cases where you have a bunch of SSDs and want to create a single large volume out of them. For spinning media, if speed is important, get it on an SSD, otherwise, any single large disk is sufficient. And in my opinion, other forms of RAID should strictly be used for high-availability applications where down-time can be business disruptive. RAID5, 6 or 10 really has no place in a workstation today or a home PC. As I've said many times, people run RAID because they can, not because it makes sense or they need it.
I guess it's somewhat of a religious argument, but I've been using SSDs now for 5 years and I can't stomach working on any computer or any task where the data is on a spinning disk. If I need to work on something that's been archived to an HD, the first thing I do is move it to my SSD.
EDIT:
Oh, and thanks for the screen shots of the pricing on the old machine.
Last edited:
I dunno man an hour of ProRes HQ is like 80GB. It's crazy to want to edit that on spinning disk media. SSDs are now cheaper than HDs were just 7 years ago (1TB for $500). There's really no excuse for anyone to be working off HDs as their primary storage medium anymore.
Having said that, HDs still have a place for archiving content or storing stuff that doesn't benefit from high speeds or frequent access like your video and audio collection or last years photos.
RAID is another matter with SSDs and very large fast high-density 3-4 TB drives, the benefits of RAID have really diminished. Even RAID0 doesn't really make sense except perhaps in cases where you have a bunch of SSDs and want to create a single large volume out of them. For spinning media, if speed is important, get it on an SSD, otherwise, any single large disk is sufficient. And in my opinion, other forms of RAID should strictly be used for high-availability applications where down-time can be business disruptive. RAID5, 6 or 10 really has no place in a workstation today or a home PC. As I've said many times, people run RAID because they can, not because it makes sense or they need it.
I guess it's somewhat of a religious argument, but I've been using SSDs now for 5 years and I can't stomach working on any computer or any task where the data is on a spinning disk. If I need to work on something that's been archived to an HD, the first thing I do is move it to my SSD.
EDIT:
Oh, and thanks for the screen shots of the pricing on the old machine.
That's how I look at it as well. I've been on SSD since 2010 and it's been amazing. I only use my HDD's to archive movies or photos or other media and if I'm working on something, the first thing I do is move it to my SSD.
What I'm thinking about my nMP is to buy a 1TB Samsung 840 Evo SSD and connect it using the Seagate Thunderbolt adapter to my Mac Pro and it'll act as the primary documents disk, the documents which need frequent access. Together with the superfast internal SSD, which I'll probably opt for a 1TB one as well, it should be sufficient for a long time. And the remaining HDD's I can either connect them again using a Seagate TB Adapter for 3.5" disks, or buy a RAID enclosure like Pegasus 2, and use in JBOD mode. (Don't need RAID0 speeds).
One thing I noticed having built my own external array out of PC parts: Quality counts for a lot.
Even if you get your TB 2 enclosure up and running, how would that compare to having the drives inside the case?
- You have another PSU you have to worry about
I've had tons of problems with the PSUs typically found in enclosures. They always give you a wattage rating but you never really know what they're putting in there. Both of the fancy firewire enclosures I've owned have had either totally defective or inferior PSUs that I've had to replace (note the ATX PSU bolted to the top of the array.. which has a broken PSU inside). We're living in an era with 80PLUS GOLD certified PSUs going for $50, and you get to settle for a cheapo PSU that could've come from an illegal sweatshop in rural china.
- Enclosures = fan nightmareMore PSUs = More heat. More heat = higher RPM. Small enclosures = small fans. Small fans = less air movement, more noise, more failures. 120mm fans are the sweet spot, but most small enclosures use noisy, whiney 40mm fans.
The old mac pro has great ventilation using large, high quality fans. It also has a decent PSU. To replicate that in an aftermarket storage solution requires big bucks. nMP apologists are considering external storage as an afterthought--like it's separate from your computer. Your external storage is part of your computer, and you should think of it that way. The cost, portability, noise, and reliability of the nMP needs to factor in the storage solution you're using. Otherwise it's just dishonest.
That $100 USB box seems enticing, but it's going to be noisy and the PSU is suspect. I would NOT trust my data to a box like that. I've done it before and regretted it, bigtime.
Last edited:
I think the pricing is about right.
Whether you agree about the inflated costs of professional components, the new Mac Pro is a very small form factor professional computer with workstation class processors, memory and professional graphics cards. At the price you're getting pretty good value if you look at the cost of the components for assembling your own custom workstation, but the difference with the new Mac Pro is that you have limited (if any) future upgradeability, which means that in terms of performance it's value is going to decrease over time unlike previous Mac Pros which can be incrementally upgraded well beyond your initial purchase.
So yeah, the comparison here is between a somewhat cheaper but similarly less capable entry level machine with tons of upgradeability, to a more sophisticated and powerful, full-featured entry level machine with almost zero upgradeability.
The other important factor as others have pointed out is that the new Mac Pro is going to cost you money for extra storage pretty much regardless of your current setup. People with huge storage requirements may already rely heavily on external enclosures, but the most common of those are Mini SAS require a PCIe card to use, so even if a Thunderbolt to PCIe enclosure is "all" you need, it's a pretty significant extra expense at current prices.
Meanwhile, people like me that were fine with whatever number of drives they could cram into the current Mac Pro are in an even worse situation; whether we upgrade to the Mac Pro or settle for a Mac Mini or iMac, we need to buy all new enclosures to recreate that storage. But Thunderbolt options are expensive, and USB options are slow and generally lacking in features (or costing just as much for one that do have desirable features).
Personally I'm looking into a fully custom RAID solution that will use USB for now, with a view to upgrading to Thunderbolt parts when prices come down, but that's a pretty extreme alternative and not likely to save a ton of money either. Plus as Slughead says; you get to pick your own components. For me I might even go for a passive PSU (there are some really efficient ones if you're willing to pay the premium) and a couple of nice and quiet 140mm or bigger fans, with a pair of 5 drive USB controllers that I can swap in future.
Whether you agree about the inflated costs of professional components, the new Mac Pro is a very small form factor professional computer with workstation class processors, memory and professional graphics cards. At the price you're getting pretty good value if you look at the cost of the components for assembling your own custom workstation, but the difference with the new Mac Pro is that you have limited (if any) future upgradeability, which means that in terms of performance it's value is going to decrease over time unlike previous Mac Pros which can be incrementally upgraded well beyond your initial purchase.
So yeah, the comparison here is between a somewhat cheaper but similarly less capable entry level machine with tons of upgradeability, to a more sophisticated and powerful, full-featured entry level machine with almost zero upgradeability.
The other important factor as others have pointed out is that the new Mac Pro is going to cost you money for extra storage pretty much regardless of your current setup. People with huge storage requirements may already rely heavily on external enclosures, but the most common of those are Mini SAS require a PCIe card to use, so even if a Thunderbolt to PCIe enclosure is "all" you need, it's a pretty significant extra expense at current prices.
Meanwhile, people like me that were fine with whatever number of drives they could cram into the current Mac Pro are in an even worse situation; whether we upgrade to the Mac Pro or settle for a Mac Mini or iMac, we need to buy all new enclosures to recreate that storage. But Thunderbolt options are expensive, and USB options are slow and generally lacking in features (or costing just as much for one that do have desirable features).
Personally I'm looking into a fully custom RAID solution that will use USB for now, with a view to upgrading to Thunderbolt parts when prices come down, but that's a pretty extreme alternative and not likely to save a ton of money either. Plus as Slughead says; you get to pick your own components. For me I might even go for a passive PSU (there are some really efficient ones if you're willing to pay the premium) and a couple of nice and quiet 140mm or bigger fans, with a pair of 5 drive USB controllers that I can swap in future.
Fair enough, but this still comes down to the question of what is one's individual use case. For example, you seemingly never look at your archives, while I've had a need within the past month to go into my database for a longitudinal survey for a 2004-present series.
EDIT: I do hope that the need for 'one big honking library' will have become obsolete with the current version of Aperture, but there's little need to change today's workflow until the current RAID is outgrown, or some other similar hardware-based factor. Frankly, I doubt that a 2TB SSD is going to sell for $100 within the next few months to justify it simply on price.
True enough, but this is also a slippery slope. For example, do you keep your archives on the cheapest storage media possible, or do you "gold plate" by keeping it on a hard drive? Or even an Enterprise-class hard drive?
FYI, I'm not criticizing/condemning, but merely pointing out that there's more to the decision than only the cost of the media...it ultimately involves the entire workflow.
Sorry, but it has already been illustrated that while SSDs have continued to improve in price, RAIDs still have a quite viable bang-for-the-buck in terms of bandwidth and capacity.
Sorry, already have one of these ... they're a useful tool for creating backups that are going to be archived off-site.
In any case, all that you're continuing to do is to assume that everyone needs the same exact workflow as you do. As soon as that assumption is incorrect, the relevance and value-added of its dependencies collapse.
FYI, a fair warning for their use: like any other external peripheral, their chipsets will have limitations on what hard drives they can support - the one that popped up a few years ago was that it couldn't support 3TB or larger, for example. Sometimes, the limitation was only in formatting as opposed to read/write, so one could format a new drive on an internal MP bay and then use it in the external. Ultimately, the external has to be removed from service and replaced with a version with a newer chipset.
Such chipset hardware support factors just become one more thing that needs to be checked out (and YA "time suck" in the land of "'Just Works' Is No Longer Included In The Box") when doing one's product selection research.
-hh
Last edited:
It's certainly not an issue of cpu pricing, as the mac pro uses a cheaper cpu than a 15" rmbp. The base model matches the specs of an E5-1620v2 We also don't know exactly what firepro card is in use. Speculations suggest 2 cut down mid range cards, but workstation cards can retail as low as $200 depending on the model. Most of the desirable ones cost several times that. It's more likely a case of low anticipated volume and high markup.
Most of the world is using laptops with very limited internal storage, and many of them deal with equally large data sets and archival needs, and somehow they have managed to solve this storage problem without complaint, and without their world coming crashing down, yet for some reason, many Mac Pro owners are either unwilling or unable to do so? It sounds petty.
If you're saying I'm stubborn... yes!
Seriously though, I'm not assuming that everyone has the same workflow, but I do think 95% (or some vast majority) of the world has the same storage requirements:
1. High performance (>250MB/s), low volume (<1TB)
2. Low performance (<250MB/s), high volume (>1TB)
There are many who believe that there is a need for a third type of storage...
3. High performance (>250MB/s), high volume (>1TB)
And I'm saying this is a nice-to-have, but not required... and adding this as a must-have requirement on your storage is unnecessarily inflating the cost of your storage solution, particularly now that you're looking at a solution outside the box. But if you truly believe you need #3, then you need to be prepared to pay for it.
As you know, I believe anyone who thinks they need #3, can get by with a combination of #1 and #2 without noticing or even suffering. The right mix of #1 and #2 may vary from workflow to workflow, but there is a good solution for nearly everyone that consists of some combination of these two storage strategies.
I'm not saying Apple's markups aren't going to be huge; but then they've always been. I doubt anyone is disputing that you could go out and get a workstation for $3,000 that out-specs the new Mac Pro in practically every way, but can you do it for a computer sitting a 6.6" foot-print that's not even 10" high, and can cool the whole thing with a single fan?
It costs a pretty big premium to pack so much horsepower into such a tiny computer, and pricing individual components can't reflect the research and development, custom manufacturing etc. etc. that goes into it.
Personally I think $3,000 is about right; sure I'm a bit disappointed they didn't keep their previous entry level price, but I wasn't exactly surprised either. The question isn't whether you can build a machine to out-spec the new Mac Pro for the same price, the question is whether you can build a better Mac Pro for the same price; just the same as asking if you can build a better iMac for less, or a better MacBook Air. Like it or not there's a bit more that goes into a Mac Pro than the individual components, even before now when it was more of a pure workstation.
I know plenty of people are disappointed at the abandoning of the pure workstation form factor, but then Apple's point is pretty clear; the Mac Pro is a professional desktop. I mean when it comes down to it, the previous Mac Pros were great, but there were always better workstations out there, even when the Mac Pros were getting new processors a little early. The question is whether the Mac Pro offers enough on top of its specs alone to make it worthwhile, and I believe so.
In terms of physical size, they're trending toward the very old direction. The G3 and G4 were nowhere near the size of a G5 or mac pro. I'm not surprised by the price hike either, but it is annoying when considering the items that must be updated along with it. I wonder if they'll be really be able to reboot the line when it starts at that price and thunderbolt peripherals start somewhat high.
Anyone who uses a laptop and relies on external storage plugged into a wall for 90% of their workflow is no longer using a portable computer. That's because that bulky storage, for all intents and purposes, becomes part of your computer when you plug it in.
It's not petty, it's a reality. A computer is still a computer, even if the components are not in the same case.
If you're using an enclosure with a loud fan and a bad PSU, that is part of your computer. Your computer is loud and unreliable.
It's not that I'm against external storage (clearly I'm not), but for many users, the 6 drive bays of the old Mac Pro were sufficient space. Forcing people to externalize 4+ hard drives that they would not have otherwise increases costs, space, noise, and risk of failure. Therefore, for many users that are forced to use enclosures they would not have with an old Mac Pro, the nMP increases cost, space, noise, and risk of failure.
Your gloating about cheapo $20 enclosures is basically saying you're comfortable with putting your data at risk and with a louder storage solution. If you want to be honest about it and directly compare the storage of the old Mac Pro to the new one, you'd refrain from using solutions that are noisy and not using 80 PLUS certified PSUs. Yes, there are tons of options (I use one), but they add significant cost to an already expensive computer.
Last edited:
Absolutely, I did kind of suspect something like this would happen for the Mac Pro eventually, but didn't expect it so soon given the huge price of Thunderbolt peripherals. I would have thought they'd wait a while longer for prices to come down a bit with laptop, Mac Mini and iMac users looking to get Thunderbolt if they can (afford it), but right now not so sure. That said, Apple have been the ones to push new standards in the past, and there are a good number of high end Thunderbolt devices out there that just about justify their cost, we just need some more choice below those options for the rest of us
I agree with this, but there is one slight advantage of external devices; you can put them somewhere else. Okay, so there's a limit to how long a USB or Thunderbolt cable you really want to use, most people will have been keeping their Mac Pros under a desk due to the size, but a new Mac Pro can happily live on top of a desk while only storage is relegated to live underneath, or even in a suitable cupboard (i.e - one where the door isn't full height so there's still airflow).slughead said:
This is exactly what I'm hoping to do, although actually my Mac Pro currently lives on my desk but I have it sideways behind a big HDTV, so not your typical set up I don't think.
Speak for yourself. I regularly spin data into the TB ranges. The biggest stuff is reading raw data files which can be huge (like say one project with some 750GB of data), then creating intermediates that range up to nearly the same size (ie that 750GB will go down to maybe 500GB). Then you compress that 500GB into maybe 250GB. Now you start reading that 250GB file and more or less cleaning it up in stages (ie you create 3-4 more versions of it that gradually srink in size down to maybe 200GB). Then you do the final analysis that creates something that's maybe 1GB.
That's one project that takes about 2 weeks and I'm often doing 3-4 at a time. Ultimately I want arhive that raw data gzipped, the first compressed intermediate and the final output. So each project has about 1TB of archive needed, and maybe 3TB of working space needed. SSDs are joke for my needs PERIOD!
Oh right. Clearly you know best to generalize for everyone that doesn't see it your way. Obviously, the sensable conclussion is that you can make a blanket statement about what these "people" should do without knowning one iota about their particular needs.
Ultimately we'll see what peoples needs are based on the sales. Lets see how this thing is doing about a year from now.
Ok, so if you had it your way with my needs, you'd have 4 of those plugged into a TB -> USB3 dock? Remember I need some USBs for keyboard/mouse, flashdrive, iPhone, backups (2) already. Add it all up and thats 10 USB devises. Sorry, no thanks. I just can't imagin that will perform well or be very stable.
Its not that this thing can't work for decent amounts of storage needs (I'll accept I'm a bit of an outlier but maybe not by a ton), the main problem is just WHY? This is a desktop for professionals. Why are you leaving the SATA I/O on the table? Its there. Apple just didn't plug anything into it.
I mean, what happens to your I/O when you're piping lots of data through the PCI via TB and trying to use the GPUs for computation, or when you plug 6 USB drives scatters through hard to know how many controllers all working at the same time? It all just sounds like a bottleneck waiting to happen for no damn reason. And it forces people into less elegant solutions with USB/TB externals flung left and right. All for what? Just to do something different? Did I miss something, did all of our desks shrink overnight?
I'll wait to see some reviews of how the performance of this thing stacks up for needs like mine to completely write it off, but right now I'm very skeptical. And everything that's been said on these pages hasn't changed that even a little.
Wow, nice assumption to lead to an insult.
Many of the laptops dealing with large datasets aren't doing anything but remotely accessing the data and keeping all computation where that data is located. Is that what we're buying nMPs for? Cuz, I can do that with $150 Chromebook.
But you need to remember this is the subset of the market that would buy a workstation computer. I think your assumption is on shaky grounds at best once we take account of that.
....not required FOR YOU.
I have payed for it. It came as 10 possible SATA connections on my Supermicro motherboard and workstation case with 8 3.5" bays + 2 5.25 Bays. All for a base price right about that of the nMP.
Hi, my name is Wally. Nice to meet you.
If you jack #1 to ~6TBs, you start to get into my ballpark.
What on earth are you doing?!So what is your storage solution like?
Based on what you're saying, you'd fill up all the drives in your Mac Pro after a month. I'm imagining a closet full of drives in static bags or in a large moving box or something.
Actually, I think I saw a photo of your storage solution... (and it's primarily external)
