Shoulf I get a retina macbook pro?

#141

I get that much, it’s the same thing I mentioned (benching it revealed the difference), what I’m asking about is if you actually found real use differences in day to day use.

The difference between an SSD and a spindle drive, even capped by IDE, is considerable enough to pick the former, and in turn hit by enough things that its topmost performance is practically never significant.

On top of that, while a benchmark will test sequential read and write in an Utopia scenario by bitsequencing a file on the same drive, how often do you actually move data from SSD to SSD (which implies having two, because locally it would just retable the file in real world scenarios), or in day to day use perform big enough sequential reads to notice the difference?

Things like loading an app, paging the memory in or out in the background and so on will not really show difference you can appreciate in terms of feeling it as a user.
Copy to or from the disk will always be limited by the spindle drives or the network you will be going through (again, unless you use dual SSDs, like having an external TBolt SSD drive, which is fairly uncommon).

I’m not debating your findings, they are well founded.
I’m saying that this statement “that cuts your disk speed in half” is too blanket-y. It cuts some of your benchmarks down 40%, the rest is likely to be between 0 and 10% impact :slight_smile:

In the use case scenario mentioned here, 3DSmax, I’ll hazard a guess that if there even was any difference, it’d be well contained in the single digit percentage realm.
I’m really not a Mac kinda guy, and I actually agree that if all you want to do is use Max and VRay you might as well go for a cheaper option with USB3 over a MBP, but I don’t think the forceful IDE mode is such a determining factor.

#142

It’s unfortunate that this wasn’t revealed earlier in the thread.

If game assets are what you will be working on primarily, then you WILL be spending ALL your time in Windows. There’s no question about this.

Many industry standard tools that help facilitate the creation of game assets which Windows users take for granted are unavailable in OSX. In fact, the lack of game oriented content creation tools has meant that even developers favor Windows over OSX for 3D based iOS games; a highly embarrassing predicament for which Apple seems only too happy to ignore given their focus on the consumer market.

I have a 2011 MBP 17 inch myself and for well over a year I have tried to find a decent workflow that would allow me to create content for games without having to flip through half a dozen different applications or resorting to bootcamp. The realization that there are no decent texture baking tools at all was especially alarming. The closest I came to was Maya with Mental Ray - hardly an ideal solution. Other applications including Lightwave3D, Modo, and Cinema4d force you to perform a bizarre array of rituals to accomplish the same end result that could be reached much quicker with 3dsmax on Windows.

Fortunately for me, I run the gamut of everything from 3D to photography to motion graphics to video and this is where OSX is at its best. So I kept my 2011 MBP. However, if your focus is aimed primarily at 3D where you’ll be working in only a few choice applications, OSX loses most of its advantages.

Had you elected to go a different route such as architectural visualization, you would have been fairly safe sticking to OSX but the opposite is certainly true with game assets.

At this point, I can’t recommend an Apple laptop. You will only be gimping yourself in the end. It was a bitter pill to swallow for me as it probably is for you but trust me when I say you will get so much more for your money with a PC laptop because Windows is where you’ll stay.

#143

Not that it goes against your point, since it’s true that game dev is, for many reasons, very windows centric, but just in case it helps:
http://www.topogun.com/

Cheap, amazing for retopo, has respectable tools for generating maps of several types from differentials built in (elevation, normal, ambocc etc.).

In case it saves you a few boot-overs :wink:

#144

Indeed.

Topogun is a fantastic tool for retopology. I have version 1.xx of the program.

I was never able to get the kind of results I had hoped for with the baking toolset offered by Topogun though. I use xNormal on a VMware install of Windows 7 and I’ve had much better luck with that. The bakes can take awhile but I don’t need to spend an awful lot of time in the application so generally a VMware fusion solution is tolerable - ditto with 3dsmax for converting .max files.

I have yet to check out Topogun 2.0.

I will say that if only Autodesk had included the Turtle renderer with Maya for OSX, then a lot of my woes would have been solved.

#145

Turtle is included in Maya 2014 on all platforms.

I get that much, it’s the same thing I mentioned (benching it revealed the difference), what I’m asking about is if you actually found real use differences in day to day use.

well, considering the OS can’t get around the limitations, then of course there would be a noticeable difference. You’re launching apps at half the speed, paging at half the speed, etc. I don’t use Boot Camp for work - I only use it for the occasional game on the road, so I can’t compare it directly but it’s obvious that subjective experience is irrelevant given that you have numbers to tell you exactly what to expect - significantly worse performance in Windows for anything file system related.

#146

Since Maya 2014?

My understanding was that in order to get Turtle you had to buy the entertainment creative suite which was PC only.

Autodesk has been very vague about Turtle for the Mac in the past.

Guess I’ll have to give it another look as its been awhile.

#147

But that is what I’m talking about.
You DO NOT launch apps at half the speed, you do not page at half the speed etc.

The raw speed of a sequential read is irrelevant when doing those things, the bottlenecks elsewhere kick in long before you get the slimmest chance to hit the IDE BW cap.

The benching showing 400MBps sequential reads are done on large files in one single sweep where the bit hopping for the cache is irrelevant and many other factors have no relevance either.

Fetching from the drive the files needed to launch an app is not that case. It takes what, 6, 7 seconds to launch a vanilla maya? And how long when you have a lot of plugins to initialize?.
In your best case scenario you have maybe a few megs to fetch (no chance to ever hit that cap) for files, and assuming you are launching it right while minimizing another hungry app a few hundred megs to page out, again not a capped scenario, but possibly a 350 vs 250 scenario.

The difference, at the end of the day, will be a 6.5s launch vs a 7s one, or in other more common cases (non vanilla) even 30s vs 31-32s. It’s NOT halving the speed. That’s exactly why I’m saying you’re blanketing this. I don’t believe for a second that you really think launching an app is 100% a matter of sequential read speed, sequential read at cap will play anywhere between 0 and 30% of that time, which means in the best case scenario you get a 15% speed up when switching to AHCI, most of the time it’s between 0 and 5%.

Launching the OS, same thing. My windows boot is about MAYBE 1s faster, and my linux boot I don’t know, because if there is a difference, it’s smaller than I can twitch the stop button on my chrono app.

#148

it’s in the default plugin list on OS X Maya 2014 install. I used it enough to see that it’s working.

ThE_JacO - if you look at the benchmark, it’s significantly slower for all types of read/writes and sizes (4K up to 1024K and sequential/random read/write etc). And that is with “allow cache effects” enabled.

#149

Let me put it this way:
I benched both, I can tell there’s a difference.
I tested both, in real case scenarios, and have a pretty decent idea of how things factor at an IT and CS levels, and the difference IS NOT that the speed of common tasks where an SSD as a primary drive (not a data drive) matters the most are between unaffected and barely perceivable. They pretty damn sure ARE NOT twice as slow.

If you keep just mentioning the benching I don’t think this conversation can go much further than this. Allow cache effects is normally an option to forcefully pipe through it to test its performance, regardless of whether the HAL interfacing components would normally choose to do, it IS NOT indicative of cache sensitive operations. It’s like turning your wipers on on a sunny day to see if they’d work in the rain :stuck_out_tongue:

#151

ya, he’s a student so he’s probably going to want to explore his options. Anyway, he’s free to get a Mac and run it in Boot Camp if he wants to try UDK or Cryengine. I just don’t recommend that as your default mode with an MBP

#152

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.