donour

iTrader: (6)

This is where I point at windows and laugh.




...although I'm not sure what OS X does either. Unix buffers all writes, so this should not be an issue. So even a blocking write will almost always return immediately if the data is small (logs are small). I mean, I probably log with the slowest machines of anybody and I don't buffer at ALL. I read from the port and do an immediate write(3).

d

EDIT: It's probably not actually the NTFS filesystem itself, but rather the disk driver can't tell the difference between a momentary disk park and a real disk park.

MalibuJack

iTrader: (5)

Generally workstation OS's do immediate writes to reduce data corruption should the machine get shut down. You can turn on write buffering, but windows gets corrupt over time if the user has a habit of not shutting down properly. This is a problem that I've seen happen all the time on older MACs from just powering them off because they DID buffer.. (Sad Mac Icon on boot)

Mr. Evo IX

iTrader: (5)

I'm pretty sure this is an issue specifically with IBM Thinkpad laptops that "brace for impact". They have motion sensors built into them which stop the harddrive to protect data from vibration / jarring / dropping. My Evo creates enough vibration to trigger the sensor which stops the harddrive which causes the logging to stop. Anyhow this behaivior can be turned off or can be dealt with in the logging software.

It is not really an OS issue but what happens if your OS X cant immediatly do a write because the HD is momentarily turned off due to shock?

donour

iTrader: (6)

Ah. But is this still a problem? NTFS is a logging filesystem.

d

donour

iTrader: (6)

All modern apple laptops have a 3d accelerometer as well that's parks the disk if sudden motion is experienced. I have not encountered this problem.

d

jfitzpat

The problem is the way that IBM implemented the feature. It really should be a filter driver in each file system stack. However, they implemented it at the physical drive interface, holding the drive 'busy' beyond the interface spec. The physical buffering scheme at the lowest level driver is based on spec compliant devices.

If anything, this is a good example of why I find Apple Windows debates so pointless. Apple is, still, fundementally a vertical, closed hardware system. Windows is fundementally a horizontal, open hardware system - which leads to lots of people, of varying skills, writing drivers, etc.

On the software side it is, again, apples and oranges. On the one hand, I can point to a long list of things that were sadly lacking on Apples (OS wise) prior to OS X. On the flip side, I could point to how much cleaner OS X is than the NT OS and Win32 Subsystem from a programmer's point of view.

But really, it is again completely different approachs. I've got Win16 apps that still run on Vista while Apple has been in a position to basically say, 'all done, throw away, new OS...'.

Having started my programming career writing BIOS ROMs and 8" Floppy Disk support for what would later become CPM (it was still "Intergalactic Digital Research" at the time - and Microsoft was the MASM company in Arizona), I'm pretty much an equal opportunity hater of all computer OS's. That said, I'd still have to say that the new Apple commercial is dead on. The one "Gadget" I really need for Vista is one that sends a FU instant message to everyone in Redmond each time the system asks me if I'm "sure" I want to do something...

-jjf

codgi

iTrader: (22)

I was actually going to jump in and say this. Its just how IBM did it that horks the OS (ran into this at the previous company whose program logged application calls). If there was some way to put OSX on an IBM it would probably jump up and down as well . In those cases there is only so much the OS can do.

Yep...something people don't understand. It also makes the testing matrix for Apple a much smaller infinity...while for windows it is always pretty much the largest infinity you can possible find (and yes all infiniti's are not the same).

Yep. Never be able to do that...so there is the "ween off process" which means remove the little bits we can once we think no one else is using them.

And when we DOS you with our responses will you come and sue us then? . 29,000 >>> 1 . The heuristics in the OS will always only be so smart...so silent killing is not the best way to deal with this stuff. Unfortunately we will have to ask the user...of course since they usually guess wrong (proven in tests long before Vista) we end back up at the "user education" phase.

TTP Engineering

iTrader: (465)

Thread is starting to drift fellas.

donour

iTrader: (6)

Isn't this a topic for logging/flash questions? I think platform discussion would fall under that.

I think it's really funny that I'm always the anointed defender of macs. For what I do, it sucks.




d

jfitzpat

That's OK, all my encounters with the Man have been pretty surreal all ready. I've seen a couch jumping demonstration, had a breakfast sausage thrown at me at the Moscone center, and very nearly soiled my pants when he sprang up in the middle of the night and demanded to know how many bugs were still open in the database (like garlic or a cross, my visitor pass seemed to protect me).

An electronic attack from hordes of loyal minions would just be par for the course...

-jjf

FastAzzEvo

iTrader: (11)

where do you change the values to stop worrying about fuel cut?

MalibuJack

iTrader: (5)

That is the Boost Limit Table.. The proper way to set it is just a bit higher than the load you might hit at different RPM ranges at your desired level of boost. If you work with multiple maps or an electronic boost controller, then you might just want to set it as high as it will go. There's also a boost cut delay timer which if you have the stock boost control in place, is where it will cut boost should it overshoot the value for the duration of time indicated (I think its in milliseconds) in the desired boost target table

my-red-rs

iTrader: (5)

Why do turners change the "Boost Control Load Offset' value from 60 to 80 or 100. What is the benifit of a larger offset number?

I am trying to update my "Boost Desired Engine Load" tables by first starting with my desired calculated Target Loads (Desired Engine Load + Load Offset) for each RPM range. Then I subtract the Load Offset to get the Desired Engine Load value, but I don't know if I should use the stock Load Offset of 60 or change the calculation to 80 or 100.

Below is a chart to illustrate the point of my question. The calculated Target Loads are highlighted in yellow (values that I want to hit). The area outlined in red is the area that I want to increase from stock. The three tables are... 1) stock table, 2) load offset of 80, and 3) load offset of 100. I purposely kept the calculated target loads the same for the non-stock tables to illustrate the impact of the different load offsets on Desired Engine Load.

Note: the calculated Target Loads listed are illustrative, and are not what I am really trying to hit. I am just try trying to understand the role of the Load Offset value.



I am a noob who is trying to eat an elephant, one bite at a time. So if I am way off base... let me know.

my-red-rs

Mr. Evo IX

iTrader: (5)

Tuners max out the values so that people get the maximum amount of boost available (with both pills installed). If you remove the pills then you have capability to overboost the system and your Boost Desired Engine Load values will need to be changed. If you use a higher value in your offset it will allow you to work in a higher range of boost. If 80 is sufficient to achieve your desired boost then use that. If you install meth and need to go higher then maybe you'll need to put 100 there. The stock value on version 13 and 15 maps is 80. What version is your stock ECU?

my-red-rs

iTrader: (5)

My car is a 2005 USDM EVO8, ECU Internal Id 96940011...

So a higher Boost Control Load Offset can put you into higher Target Load values without having to dramatically change the Desired Engine Load? Correct?