NOD32 3.0 here: is it not enough to exclude from realtime protection specific folders and all of the files inside them (trading applications folders such as C:\TT)?
Multicore processor question
- Thread starter BoyBrutus
- Start date
NOD32 3.0 here: is it not enough to exclude from realtime protection specific folders and all of the files inside them (trading applications folders such as C:\TT)?
90% of the services running in any Windows install are never used. They are in there "in case" the user wants to use them. Great examples are stupid shit like Messenger, Games, Blah, Blah, Blah. This shit bogs down our systems and only causes us more and more latency and problems. Every time Billy Goat Gates comes up with another trick of the trade, they load it into our startups and screw us for life unless or until we discover how to hack it out of there.
Why are we always fighting our systems to get what we want, which is a computer running what and only what we want to do with the system which we bought and paid for?
Why are we always fighting our systems to get what we want, which is a computer running what and only what we want to do with the system which we bought and paid for?
google: unnecessary services
Most programs won't use multiple cores. Even those that do, usually don't benefit from it because in most cases the bottleneck is hard disk / memory access, not the CPU. Even if Task Manager shows 100% CPU, it probably spends most time waiting for memory fetches, not processing data. The kind of tasks that really benefit are those that work with small data sets, smaller than the size of L2 (preferably L1) cache, and most interesting tasks do not fit into this category.
I am working on multithreaded applications, its rather tricky. You should serialize disk access, serialize memory access, sub-divide tasks, synchronize efficiently, ...
I am working on multithreaded applications, its rather tricky. You should serialize disk access, serialize memory access, sub-divide tasks, synchronize efficiently, ...
I am not aware of any mainstream anti-virus program that monitors the network traffic - the performance issues would make such a task impossible. Dedicated IDS/IDP appliances cost thousands of dollars to accomplish that goal on servers that do nothing but monitor network traffic. Most PC's now come with gig copper NICs - do you really think that mcafee/symantec's desktop clients can scan gig speeds on the fly?
All AV programs I am aware of rely on the fact that you download something from the internet to the drive before executing, thus they scan files, not streaming network data.
Try it for yourself, bring up task manager and then go to a "speed test" web site and launch the test. Watch which processes get busy and which don't.
Anti-Virus exclusions are defined by disk directories/files, not programs. If you don't want anything related to program XYZ scanned by McAfee then just exclude \Program Files\XYZ\... and that should do the trick unless the program stores data somewhere else on your system.
In terms of disabling unneeded or seldom used services - I used to play that game when memory was more expensive but to be honest its seldom worth it anymore. Memory is dirt cheap - buy an extra 2GB of ram and that will more than cover all the crappy services - note that most of them don't consume significant cpu so their impact is during system startup (when they init) and their memory footprint. The only service I disable are the active ones, like the Windows indexing service .
Another poster talked about 90% of services being unnecessary and then gave examples like "Games" - this is a case of not knowing the differences between services and programs. Most installed programs that you never run don't consume anything but disk space. If you want to uninstall them to clean up your start menu and free some disk space fine, but it wont make your machine run any faster.
Some of what you say, while not wrong, is slightly misleading imo.
The best book I have ever seen on the subject is this and should be owned and read by every programmer:
http://www.amazon.com/Art-Multiproc...=sr_1_1?ie=UTF8&s=books&qid=1224943825&sr=8-1
nitro
The best book I have ever seen on the subject is this and should be owned and read by every programmer:
http://www.amazon.com/Art-Multiproc...=sr_1_1?ie=UTF8&s=books&qid=1224943825&sr=8-1
nitro
Yes, most software now needs to be rewritten to support multi-core processors optimally.
In fact, Intel has just released a software kit to do just that....funny how it has come about 2 years after the hardware was introduced. Once again, software tech lags hardware tech.
http://www.intel.com/go/parallel
Of course, we all know which one is more important, right ?
In fact, Intel has just released a software kit to do just that....funny how it has come about 2 years after the hardware was introduced. Once again, software tech lags hardware tech.
http://www.intel.com/go/parallel
Of course, we all know which one is more important, right ?