I've been poking around the settings, you do not need 4GB ram to run vertminer getting insane amount of hw error. I tried on my gaming rig with 16gb ram still getting hardware error with incorrect setting. imo the ram dun matter that much but you will still need a min of 2gb for scyrpt mining.
1. check if you are mining from a p2pool. most likely your card is trying to mine blocks with too low difficulty.
My worker is getting a lot of hardware errors
Because of the short blocktimes on the p2pool sharechain, your workers need to receive much lower difficulty work than they would in a traditional pool. They must finish their work and submit it before it goes stale - on the main blockchain, this would happen every time a block was generated (every 2.5 minutes on average), wheras on p2pool a new block on the sharechain happens every 10 seconds on average. Very low difficulty work often causes hardware errors. You can prevent this by telling p2pool the minimum difficulty your worker requires. You do this by appending a string to the end of your username (your Vertcoin address) in your miners startup command. The formula for calculating what to append is:
Vertcoin Address + 0.00000116 * Hash Rate in KH/s
So, if your hashrate was 350KH/s, you would calculate:
0.00000116 * 350 = 0.000406
and if your Vertcoin address was ViPBVm4sbXT38h9J23GkAWfNKiuPwVUYQX then you would use
ViPBVm4sbXT38h9J23GkAWfNKiuPwVUYQX+0.000406
as your p2pool username.
You don't want to set this value too high, or you will increase your stale shares - it sets the minimum difficulty for the shares you receive, and if you set it too high you will receive a large percentage of shares that are too difficulty for your worker to complete before the next block is generated, so the work will go stale. Setting this parameter is a balancing act between submitting work fast enough to avoid stales, and obtaining work that is difficult enough that it doesn't produce HW errors, or consume unnecessary bandwidth (if you have a very low minimum difficulty, your worker has to ask the pool for work too often, consuming extra bandwidth unnecessarily).
Here are some suggested values for various hashrates:
KHash/s Difficulty
50 +0.00005821
100 +0.00011641
250 +0.00029103
500 +0.00058207
750 +0.00087310
1000 +0.00116414
please refer to the following guide for details
https://vtcpool.co.uk/p2pool-guide.html2. to find the right setting balance btw concurrency thread and intensity. thread-concurrency works in a 64 block
for example if you put 8256 you will mine at concurrency block of 8256-64 = 8192
in order to mine at thread-concurrency of 8256 set your thread-concurrency at anything 8256 + 0~64 it will automatically round down to the next nearest 64 block.
you can check the file generated to see which thread-concurrency it is running at by checking your folder for "scrypt140121Tahitiglg2tc
8256w256l4.bin"
3. imo thread-concurrency uses ram on you gpu, the higher you set , the more gpu ram will be used to some point miner will tell you that you do not have enough ram
4. higher thread-concurrency doesn't seems to affect much on the overall kh/s but intensity does. you will need to find a thread-concurrency that allows you to go for highest intensity without hardware error. unfortunately at this point of time, from what I read, same card may not work with same thread-concurrency / intensity combination. you will need to manually poke the right value for your card.
my config setting is
"intensity" : "18",
"worksize" : "256",
"lookup-gap" : "2",
"gpu-threads" : "1",
"expiry" : "30",
"scan-time" : "5",
"thread-concurrency" : "6000",
"temp-cutoff" : "88"
5850 sapphire running at 830/875mhz oc via trixx, higher ram/gpu speed will affect the kh/s but to some point the increase in ram/gpu speed will not benefit much as compared to heat increase. find the balance to get the most out with least heat for a longer lasting hardware.
lastly a screenshot of my 2gb pc running without hardware error. happy tuning and share your finding here.
http://i.imgur.com/hi2a8Hk.png 
"--no-adl"
