Recommendations on New Laptop/Desktop

  • 1
  • Question
  • Updated 5 months ago
I've used Mindmanager for what seems like 2 decades now. I have extremely complex project maps that may have 5000+ topics. I will have to buy a new desktop or laptop soon, and I am trying to figure out if there is a consensus within this community about the computer characteristics that will be most important in running large maps seamlessly -- including moving the map by dragging on the canvas, smartrules, searches, filters, etc.

In your view, is it mostly about CPU and RAM?  If CPU, specific views re # of cores, caches, etc?  Obviously will go with an SSD. In my experience, a dedicated graphics card in a high-end laptop had no impact compared to the built-in gpu.  Any thoughts would be appreciated.
Photo of jerryk

jerryk

  • 11 Posts
  • 1 Reply Like

Posted 5 months ago

  • 1
Photo of Detlef Huß

Detlef Huß

  • 29 Posts
  • 11 Reply Likes
Moin Jerryk,

when coming from a mechanical hard drive, using a ssd makes a difference!

Nevertheless, I would guess myself as a power user too as you are. The PC hardware has never been a bottleneck for me when running the MindManager. Therefore you may put more focus on monitor and printer.

I've made it an experience for myself meanwhile to refrain from buying new systems, instead I'm going for used ones. My personal preference is MS Surface, which can be obtained for only some hundred bucks.

In essence, my personal bottleneck is to stay ahead with the increasing functionality of the MindManager. :-)

Regards,
Detlef
Photo of Alf Christophersen

Alf Christophersen

  • 81 Posts
  • 11 Reply Likes
As Windows is a multiuser/multitasking OS, Windows several times a second swaps out your working threads to virtual memory, an event that is less often hapening if your CPU has more kernels
But the bgger RAM, the more threads can be pushed into the non-kernel RAM.
If not free memory, your works is swapped to the drive so then slow spinning drive versus SSD comes into question.
If estriction yourself to work with few programs open all the time, the leaner computer is needed and money sspent :-D

So it comes to economy, what is most important, as fast computer as possible (and maps generated goes fast, but costly or would the money lost on just waiting for 5 hours generating the map cost less than the difference in money spent on buying the fastest combination on market :-D  That's your decision :-)
But in some cases a seemingly fast computer can be very slow if the computer start lots of software maybe not needed.

Photo of Alex Gooding

Alex Gooding, Champion

  • 1274 Posts
  • 354 Reply Likes
I don't have any specific recommendations, except to say that I would probably prioritise RAM over CPU or GPU speed, provided these are reasonably fast. I would also get an SSD, and as Alf says try not to have too many other programs open at the same time. 

I do have two other points I'd like to make.

The first is that I get very nervous when people talk about having very large maps with thousands of topics, as you are putting all of your eggs in one basket if the file fails for any reason. I've had large maps (but not as large as yours) suddenly become unstable once they reached a certain size and complexity. Obviously as a power user you know what you're doing and I appreciate that there may not be an alternative, but the advent of rolled up dashboard maps in recent versions of MM makes it possible to break large maps into smaller ones linked by a dashboard map. And if you must keep everything in a single file this size, save the map frequently and back it up.

Second, the features added or expanded in recent maps such as SmartRules and improved formulas are great but they do seem to slow things down in very large maps. I think this is because every SmartRule and formula gets recalculated every time you add anything to the map. It is possible to turn off the former until you need them but not (as far as I know) the latter, so use these features and especially formulas selectively in large maps. Again, it might be possible to apply them in a separate dashboard map linked to the larger one.

(Edited)
Photo of jerryk

jerryk

  • 11 Posts
  • 1 Reply Like
Thanks for all the replies, esp. Alex's.  All good points he makes re back ups, making smaller maps, etc.  Rollups turn out to be of limited value in my use case unfortunately. (Sidebar: I have also tried mightily to have Mindjet allow color highlights to roll up to the parent via smart rule; that would make seeing collapsed items of importance so much easier. Unfortunately, I haven't been able to get them to do that.) 

So, it's all about CPU and RAM given the way the code is written.  GPU is irrelevant.  Re CPU, I can't tell whether # of cores matter (don't know what's multi-threaded, etc.), various L2/L3 caches, etc. etc.  And I haven't stayed current on chipsets for the past few years.

Here's my basic surmise:

1. SSD is a given.
2. GPU is irrelevant.
3. 16GB RAM is plenty (probably 8MB will suffice, even for somewhat heavy multitasking). Rarely does Mindmanager take more than 1GB of memory. 32GB will have no impact.
4. CPU is what I'm curious about..... Anyone out there with more anecdotal evidence?

Many thanks.


Photo of Alf Christophersen

Alf Christophersen

  • 81 Posts
  • 11 Reply Likes
CPU kernels and RAM is connected.
No program is allowed to load whole program into memory. Windows in itself occupies most of the RAM, but only a cetain amounts of 'pages'. If there are active code waiing in queue for a kernel thread, every time slice the most imported get access after operating system has kicked your code out in shared memory, and if not space for those, the code is written to disk. The more kernels and the more RAM available for for pages, the less chances your code is kicked to disk. But, the more data program generates in memory, the bigger chance your eunning code will be sent to disk and if you haven't prepared for that by generating a swap area on disk operate system must allocate memory on disk as small fragments. And creating such virtual pages is seriously slow. So the less RAM the and the more sleeping prgorams operative system use or you have opened yourself, the bigger chance at at end computer is just swapping because before next tie slice is released another process need the CPU.

So in practice 32 Gb is faster. But, you may wait with hat investment. Check primarily how much computer is scaled for.
Look for max 64 GB or 128 GB.
Download Sysinternals Suite from Microsoft (it's free) and exchange the taskmanager and use Process Explorer (ProcExp) as taskmanager.
THere is a few videos on YouTube searching for Sysinternals videos. Some by author but there is also one about 5 tweaking tools of which Process explorer is one of these. Also ProcessMonitor is nice tool .
By studying Procexp with your map the program will tell you if swap area is a limit or what, and also how many threads are active and context swaps
One of the ressources is a 2 hour long run through about finding areas about how programs usage of the 'sleeping time' btw. task managers sampling on who is occypiying your kernels.
Photo of Alex Gooding

Alex Gooding, Champion

  • 1274 Posts
  • 354 Reply Likes
Re your sidebar, as follows: "Sidebar: I have also tried mightily to have Mindjet allow color highlights to roll up to the parent via smart rule; that would make seeing collapsed items of importance so much easier. Unfortunately, I haven't been able to get them to do that."

This is actually possible in MM2020 if you combine SmartRules and formulas, and there are two ways (at least) of doing it, depending on what exactly you are trying to highlight. Unfortunately it doesn't seem possible to roll up subtopic formatting such as fill colour directly, but virtually anything that you might be using to trigger formatting at the subtopic level can also be applied to parent topics.

And apologies - this comment is a little long. 

Task Attributes: If you want to highlight task attributes both in subtopics and their parents such as progress (or lack of it), priority, duration, cost etc, there is an extensive range of task information available as topic properties to both formulas and SmartRules (and others such as On Track Status and Days to Task Due are available as topic properties through the MAP add-in) . 

Flagging an issue with any of these at the subtopic level is easy enough - you just need to write a SmartRule with the relevant attribute(s) as a trigger and then the apply a fill colour or other highlight as the effect - say, for example, if there has been less than 25% progress you could change the topic colour to red. 

To flag the existence of slow progress subtopics (to take the same example) in the parent you need to use a formula in conjunction with a SmartRule. The great thing about formulas is that you have an extensive set of options available in terms of topic range, including children and descendants, so you can set up a topic property on a parent topic that references topic properties. 

The major limitation however is that MindManager formulas have a limited range of functions, ie, the arithmetic operators and the functions Sum, Average, Min, Max and Count - so you have to work within these constraints. You also need to think about how the attribute would be handled numerically by the relevant topic properties - so for example, 100% progress is stored as 1, and 25% as 0.25 (you can reformat the topic properties to percentages if you wish) - and then which of the functions can best be used, and how, in conjunction with a SmartRule.

For example, if you want the parent topic to change colour to red if any of the subtopics has less than 25% progress, the formula could be as simple as:

 [Flag]=MIN(Children.Progress). 

This will create the calculated topic property "Flag". The associated SmartRule would then be something like this:

Trigger - Properties: Flag: Is Less Than or Equal To: .25 (or 25, if you change the Flag topic property type to percentage)

Effect - Fill Color: Red.

The parent topic should now be coloured red even if only one subtopic is showing slow progress.

Non-task attributes/properties: If you want the parent topic to highlight a factor that isn't specifically task-related which is affecting certain subtopics it's a little more complicated but the same principles apply with one additional step. 

If, say, you want to highlight a parent topic which has any subtopic with a specific tag you would first add a topic property to all subtopics and then use a SmartRule to assign a fill colour to the relevant topic(s) as well as a number to the topic property (which I'll call Tag#).

From that point on it's similar to the above, but the formula is a little different as you are referring to the topic property you created, and you probably would use MAX instead:

[Flag]=MAX(Children.[ Tag#]

- and the SmartRule would be:

Trigger - Properties: Flag: Is  Greater Than or Equal To : 1

Effect - Fill Color: Red.

The parent topic should now be coloured red if one or more subtopics has the matching tag.
(Edited)
Photo of jerryk

jerryk

  • 11 Posts
  • 1 Reply Like
Very helpful.  Unfortunately,I'm still in MM 2019 (didn't see the benefits of going to 2020, and I don't think the Progress field is available to me as a formula option although it does appear in smartrules).  I'm also concerned re formulas and smartrules scaling on a very large complex map.  In any event, I'll see give it a try when I find a moment. 

/on_soapbox

Larger point: I hope that fellow members can see the missed opportunity here.  The beauty of maps for project management that don't follow some strict GANNT methodology is that you can add tasks into project clumps with reference information and target dates in free form. At any time, you can zoom out to the big picture simply by collapsing topics.

But doing so means that something important, such as an overdue task (16 levels deep), can be hidden. If we could roll up some visual feature (i.e. color) to the parent, and the parent always picks the "hottest" color of its descendents if there's a conflict, then a visual cue would always be available -- regardless of the state of the map's open/collapsed state.  

Seems such an elegant solution without asking folks to program formulas, etc. Just think of it as a visual "roll-up" that functions more subtly than the extant task "rollup" feature, which adds a whole mess of task icons and metadata on all branches. I find this unhelpful since many of these branches are not "tasks" and instead thoughts, reference material, brainstorms, etc. 

Remember, the beauty of Mindmanager is freefrom and dynamic clumping of very different kinds of items. 


/off_soapbox

Again, thanks.
(Edited)
Photo of Alf Christophersen

Alf Christophersen

  • 81 Posts
  • 11 Reply Likes
Why bidirectiinonal synchromization is needed is because in many companies has Microsoft office as a standard for all works, but many dislike setting up the needed settings to start work and IT personals want people rather use e-mail for discussions.
While using Mindmanager, setup and describbing project from brain storm theough organisating all ideas in a map which then could start as a start point of a new project and then to satisfythe Office guys the plan is exported to their tools.

Having bidirectional synchronizing the most positive sides of each tools may be possible to use.
So hope in future bidriectional synchronisation is possible.