Jump to content
Search In
  • More options...
Find results that contain...
Find results in...

mariushm

Member
  • Content Count

    11,111
  • Joined

  • Last visited

Reputation Activity

  1. Informative
    mariushm got a reaction from Mark Kaine in How to Lose 100TB of Data: A Guide to My $2000 Mistake   
    If you just shoved 12v on the 5v input, there's a decent chance you can repair the drives.
    Often there's a tiny component by the power input which acts as a fuse and blows up... you can simply desolder that component and replace it  or just not install anything (if it's a zener diode) or just have a blob of solder or a wire there to override it.
     
    That's... 5 minutes for each drive with a soldering iron to fix.
     
    Worst case scenario, 12v got to the actual chips blowing them up, in which case you could in theory buy some faulty drives and transplant the chips to your drives.
     
    Someone like Louis Rossman could probably make you a deal and repair them all for a big discount, since you know the reason and most likely all drives will have same failure reason so once one is diagnosed repairing the others won't take a lot of diagnostic time.
    Would probably cost you at least $3-500 if they don't have to buy other drives to harvest components from other drives.
     
    ps can you take as clear as possible pictures of your circuit boards on the broken drives?  if they're the same, 1-2 pictures is enough.
  2. Informative
    mariushm got a reaction from Stahlmann in Possible to add "VESA DSC" through firmware update?   
    There has to be hardware support in the chip that decodes the digital signal from your video card. 
    It is possible for a chip to have that functionality but the firmware didn't implement it yet, but often it's not implemented because it's buggy or something like that, requiring a new hardware revision of the chip to work right.
     
  3. Agree
    mariushm got a reaction from Kilrah in Can I harm my Video Card (Nvidia 1660 something-or-other 6GB) by putting Kaptop tape on the bare solder to stop short circuits?   
    wall of text ... not gonna read that.
     
    There's a small risk of chips overheating, if you cover them in kapton tape. The package surface no longer gets a bit of airflow over it, helping with the cooling ... but if those chips are so hot to the point that lack of natural air convection would kill them, they needed heatsinks in the first place.
     
  4. Agree
    mariushm got a reaction from Levent in Can I harm my Video Card (Nvidia 1660 something-or-other 6GB) by putting Kaptop tape on the bare solder to stop short circuits?   
    wall of text ... not gonna read that.
     
    There's a small risk of chips overheating, if you cover them in kapton tape. The package surface no longer gets a bit of airflow over it, helping with the cooling ... but if those chips are so hot to the point that lack of natural air convection would kill them, they needed heatsinks in the first place.
     
  5. Like
    mariushm got a reaction from Spotty in How to Lose 100TB of Data: A Guide to My $2000 Mistake   
    If you just shoved 12v on the 5v input, there's a decent chance you can repair the drives.
    Often there's a tiny component by the power input which acts as a fuse and blows up... you can simply desolder that component and replace it  or just not install anything (if it's a zener diode) or just have a blob of solder or a wire there to override it.
     
    That's... 5 minutes for each drive with a soldering iron to fix.
     
    Worst case scenario, 12v got to the actual chips blowing them up, in which case you could in theory buy some faulty drives and transplant the chips to your drives.
     
    Someone like Louis Rossman could probably make you a deal and repair them all for a big discount, since you know the reason and most likely all drives will have same failure reason so once one is diagnosed repairing the others won't take a lot of diagnostic time.
    Would probably cost you at least $3-500 if they don't have to buy other drives to harvest components from other drives.
     
    ps can you take as clear as possible pictures of your circuit boards on the broken drives?  if they're the same, 1-2 pictures is enough.
  6. Agree
    mariushm got a reaction from Aero_db in How to Lose 100TB of Data: A Guide to My $2000 Mistake   
    If you just shoved 12v on the 5v input, there's a decent chance you can repair the drives.
    Often there's a tiny component by the power input which acts as a fuse and blows up... you can simply desolder that component and replace it  or just not install anything (if it's a zener diode) or just have a blob of solder or a wire there to override it.
     
    That's... 5 minutes for each drive with a soldering iron to fix.
     
    Worst case scenario, 12v got to the actual chips blowing them up, in which case you could in theory buy some faulty drives and transplant the chips to your drives.
     
    Someone like Louis Rossman could probably make you a deal and repair them all for a big discount, since you know the reason and most likely all drives will have same failure reason so once one is diagnosed repairing the others won't take a lot of diagnostic time.
    Would probably cost you at least $3-500 if they don't have to buy other drives to harvest components from other drives.
     
    ps can you take as clear as possible pictures of your circuit boards on the broken drives?  if they're the same, 1-2 pictures is enough.
  7. Agree
    mariushm got a reaction from DailyProcrastinator in Is there any way to combine two graphics cards vram?   
    No, it's not possible.  Unless the software can detect both cards and run some things on one card and other things on the second card.
     
     
  8. Informative
    mariushm got a reaction from Dumb guy in My cable is gonna melt :( plz help   
    The video card may or may not pool the power together.
    The card has two main dc-dc converters, one that powers the gpu chip (let's say 100 watts) and one that powers the ram on the video card (let's say 10-20 watts).
    Each dc-dc converter  can have one or more phases - think of it like cylinders on a car, and each phase can be powered from various sources .. think of power like squirts of gasoline being injected into the cylinders.
     
    So in the case of your video card, the gpu dc-dc converter may have 5 phases, and 2 out of 5 phases may be connected to the pci-e slot, and the other 3 are connected to the pci-e 6 pin connector.
    So those 100 watts the gpu need are split into 40w to slot, and 60w to pci-e 6pin...
     
    Anyway ... each molex connector is rated for up to 5A of current .. or 5x12v = 60 watts. So if you have a 2 x molex -> 1 x 6-8 pin pci-e adapter, it should be fairly safe to use, as those 65w are split between 2 molex connectors.
     
  9. Agree
    mariushm got a reaction from Ash_Kechummm in Recording Codec Help/ Editing   
    AVI YV12  and AVI RGB24 are uncompressed formats.
     
    If I remember correctly, YV12 makes groups of 4 pixels, and uses 6 bytes to store the information about that: 1 byte for brightness information for each pixel so 4 bytes in total,  then 2 bytes with color information (the color info is shared between the 4 pixels, so you lose some color information, it's not 100% lossless)
     
    So a 1920x1080 picture will use 1920x1080 / 4 (groups of 4 pixels) * 6 (bytes for each group of pixels) =  3,110,400 bytes per frame ... 60 frames per second  = 60 x 3,110,400 = 186,624,000 bytes per second = 178 MB per second.
    So an hour would take 3600 seconds x 178 MB = 640 GB ...
     
    RGB24 stores 3 bytes per pixel, red, green and blue for each pixel, so it will use more bytes per frame.
     
    MPEG1 is very fast but is bad for near lossless, as it was not designed for that ... think of it like a codec designed to be like when you save JPG pictures (to lose some detail while still giving good quality), but because it was designed decades ago when computers were much slower, it's simply not designed to retain a lot of quality,  think of it like you have a quality slider for quality between 1% and 100% but mpeg's maximum quality is something like 70%, you can't do better. 
     
    XVID is h263 or mpeg4 asp, what was before mpeg4 avc / h264, and it's much better than mpeg1 and good for lossy compression, but not optimized for more than SD content (less than 1280x720)... it will still work well with HD but you get better results with h264.
    XVID has a lossy mode which has better quality than mpeg1 and also has a "near lossless" mode which can compress your frames in an almost lossless mode, basically visually not noticeable. You can use its settings to trade cpu usage for bigger file sizes but the cpu usage is still quite a bit high.
    Problem with xvid / mpeg4 asp is that it's not that well supported by editors.
     
    h264 is your best choice especially if you use OBS and use hardware encoder in your video card to capture, because you don't use any cpu. You can configure the hardware encoder to make bigger files retaining as much quality as possible - if you go with a high bitrate like let's say 60-100 mbps (around 1 GB per minute of video) you'll probably get the maximum quality the hardware encoder can offer.  It won't be lossless, but considering that Youtube will recompress your video anyway, and that you'll send a compressed video to Youtube in the first place, it's debatable if you really need near lossless or lossless video as source. 
     
    The software encoder x264 can be used in OBS and this uses strictly the cpu to encode the video so it will use cpu usage, but it's more versatile. You can configure it to be true lossless (but the format used is not so well supported by editors, so not recommended), it can be near lossless (visually lossless, but probably around half the size of uncompressed video) or you can configure the amount of quality you want to keep. Also has loads of parameters that you can tweak to trade cpu usage for bigger file sizes.
     
  10. Like
    mariushm got a reaction from Electronics Wizardy in Recording Codec Help/ Editing   
    AVI is not a codec, it's a container like MP4 , MKV
    In the container, you can use various codecs to compress the video content and the audio content. Your file size was super large because Bandicam probably either chose no codec (lossess, lik saving a series of bmp pictures) or chose a default codec with very small compression.
    There are lossless codecs which can do good compression, probably one of the best these days if MagicYUV : https://www.magicyuv.com/ - the FREE version is fine for your needs. Other lossless codecs that are well known are HuffYUV or Lagarith.  So you have to go in Bandicam, choose AVI and then choose the codec and its options.
     
     
    These codecs will use CPU to compress the captured data, so the games will have less CPU for themselves, so you'll get less FPS.
     
    I'd suggest giving up Bandicam and going with OBS. 
    You have the choice of sticking with software compression (using x264 to compress captured stuff to h264 with your own settings) or you can use hardware compression and choose some preset that's very high quality, like 40-60 mbps, resulting in a bit lossy but quite high quality captured stuff. 
    With the software option, you can configure x264 to use very little cpu but produce high size files, but would still be smaller than lossless avi files.  For example, you may get around 100 GB for one hour of 1080p content instead of 250 GB per hour.
     
    If you're interested about that, I can give you more details, but basically it's just using proper custom parameters for the x264 software encoder in the obs configuration.
     
     
  11. Agree
    mariushm got a reaction from leadeater in Windows 10 bug corrupts your hard drive on seeing this file's icon.   
    You can delete those using UNC paths .... something like  \\?\\c:\folder\filename.extension
     
    It's also useful when you need to rename a file because you can't delete it otherwise due to the total file path being bigger than around 250 characters.
     
     
  12. Agree
    mariushm got a reaction from AbydosOne in Weird compression idea... Probably dumb but...   
    But how do you STORE 1601 in a  file?  You have to use a number of bits for every digit in 1601 so that you can decode and recreate original content
     
    Let's say your algorithm is this:
    - [ 3 bits = run length - 1] 
    - [ 1 bits = the bit]
     
    So your 1601 would be stored in a file in 12 bits  :  [0 0 0 1 ] [ 1 0 1 0] [0 0 0 1] 
    1 :  store 0 in binary, store 1
    60 :  store 5 in binary (101), store 0
    1 : store 0 in binary, store 1
     
    You compressed your 8 bits in 12 bits ... you expanded it, you didn't compress it.
     
    Your original input was 8 bits ... which you visually show in binary as 10000001 (64 bits) and you can compress to 1601 (32 bits without clever bit packing) and by making some rules and restrictions (your algorithm) you can shrink it to 12 bits.
     
    But your algorithm only makes sense from a compression point of view, if you have a lot of  continuous bits of same state (runs of 1s and 0s), because you always use 4 bits to compress a sequence, which inflates everything.
    IF you read a file and convert each byte to bits, and count how many runs of 1s and 0s you have, you will find that rarely you will have long sequences of 0s and 1s to shrink this way ... you need to find patterns of 1s and 0s, not just 1s or just 0s.
     
  13. Like
    mariushm got a reaction from Padblaze in [psucultists] PSU Tier List   
    I'd change the Tier D - Not Recommended to some different wording.  Some people may read it as "Trash, avoid them" and they'll go for some local no-name crap instead which will be even worse.
    There's quite a few solid power supplies in that Tier D category, which could be considered "high end" in some regions where there's not much selection or options to buy online.
     
    It would make more sense to have the list as an excel/google docs sheet because it would allow for some comments columns
     
    I'd add a comment to psus that are rated for less than 40c ambient temperature ex EVGA N1 being rated 400w @ 25c or older Corsair VS rated for 30c (current models are 40c)
    I'd also consider warranty - for example should be noted any psu that has less than 5 years warranty if it's in tier B... there's power supplies with 5y or more years warranty in Tier C
     
    The list should use a better font, maybe something sans-serif, makes it more readable.
    Even using source code snippets would make it more readable, here's an example :
     
    Aerocool Project 7 Asus ROG Thor BeQuiet SFX-L Cooler Master V Semi-Modular, Masterwatt Maker MiJ, V-Gold 2013/2018, V-platinum 2013 Corsair RMx, AX 2013, RM 2019, HX Gold 750w+, TXm Gold, Vengeance Silver, SF, AX 2018 Cougar GX-F Aurum, GX-F Enermax RG, GX EVGA PS, G/P/T2 Enermax Platimax FSP Aurum PT/Xilencer/Pro, Hydro G/X/PT/PTM, Dagger SFX Gamdias Astrape P1 Kolink Continuum Mistel MX NZXT E Rosewill Capstone, Lightning, Silent Night, Photon Sama Armor Gold/Platinum, Forza Titanium Seasonic Focus GX/GM/PX, Prime, X/XP2/XP3, Snow Silent Super Flower Golden Green, Leadex I/II/III Thermaltake Toughpower Grand Gold/Platinum, iRGB Plus, DPS Gold XFX Pro Black Edition Limited, XTS 1000w, Pro Black FM 750w+  
  14. Agree
    mariushm got a reaction from Hackentosher in Cheap soldering station or iron   
    If you want something to last you a long time, get a soldering station with adjustable temperature control.  You want one with an actual temperature sensor in the tip and proper temperature control, not those sticks or cheap stations that just control the amount of power sent to the tip, without actually measuring the temperature.
     
    Hakko 936 clones and modernized designs of that ancient soldering station are fine, you'll find replacement tips easily and spare parts.
    As a beginner you'll rarely change tips so don't be so concerned about it.
     
    In US, Newark  (it's Farnell in other regions) rebrands some soldering stations under their Tenma brand and they're decent and cheap:
    21-19800 - Tenma - Soldering Station, 24 V, 900°F
    21-10115 - Tenma - Soldering Station, 110 VAC, 896 °F
    21-19750 - Tenma - Temperature Controlled Soldering Station
     
    The first link is cheap but only 35w, which means it will take longer to get up to temperature, and will work harder to keep the tip to configured temperature, compared to 60-75w stations. For a beginner it would still be fine.
     
    The Yihua 939d+ from Amazon seems OK, and as it's shipped from Amazon it should get to you in reasonable time.
    A plus with the Newark models... while a bit more expensive... you do get good warranty and you can also add good quality solder and flux to your order.
     
    Good solder is a must, aim for lead based solders, you should still be able to buy them in US, here in Europe only companies are allowed to buy leaded solders these days. Ideal would be 63/37 solder, the 60/40
     
    Liquid flux really really helps, even if solder actually has flux inside, adding a drop of flux before soldering or desoldering something really helps and makes soldering much easier.
     
  15. Agree
    mariushm got a reaction from tridy in [Memtest86] Max Contiguous Errors: 1   
    Also test it in another slot 
    Also test it at a bit lower frequency - 3200 Mhz instead of 3600 Mhz for example
     
    If that problem keeps showing up, I'd return it.
  16. Informative
    mariushm got a reaction from hirusha.adikari in One Hdmi Port Two Monitors (Different Inage)?   
    For up to 1920x1200 60 Hz, you can use a simple passive  DVI - HDMI adapter or you can buy a DVI-HDMI cable  ... the adapter is probably a better choice, as the cable would be more expensive than a HDMI cable. 
     
    Example  DVI Male - HDMI Female adapter : https://www.amazon.com/Rankie-Adapter-2-Pack-Gold-Plated-Converter/dp/B00ZMVGTA2/
     
    Exampel DVI Male - HDMI male cable : https://www.amazon.com/Rankie-Cable-Rated-Speed-Bi-Directional/dp/B00YOSA85Q/
  17. Like
    mariushm got a reaction from Aereldor in Best approach for getting 4k@120hz from PC to an LG CX from a distance of 15m   
    You're both incorrect. 
     
    The cable is still cable, the copper wires have resistance and therefore voltage drop. The longer the cable, the higher the losses, which makes individual bits harder to decode, makes the signal more "fuzzy"... at that high bandwidth each bit "takes" a much shorter time on the cable, and there's higher chances of bits to be corrupted and that can cause drops in HDCP encryption or bad blocks in the image.
     
    At longer lengths and high resolutions and refresh rates (more bandwidth), the cable quality matters more. 
     
    They worked fine because your resolution and refresh rate was long  small, as was the distance. your cable was probably 1.8-3m (10ft) long, not 15 meters. 
    Also, 4K 120 hz is 2 x 4K 60 hz bandwidth, or ~ 8 x 1080p 60 Hz bandwidth ... so quite a difference. 
     
    Optical fiber cable would be the ideal choice, as long as the chip in the cable can actually recognize and work with 4K AND 120 Hz. Some may be picky about it. 
    Regular hdmi cables with signal amplifier chips in the middle should be cheaper and should work fine. 
     
  18. Agree
    mariushm got a reaction from Naijin in WD Black SN750 (250GB) or TEAMGROUP MP34 (256GB)   
    I'd trust the WD drive a bit more. It's a good model with good controller and good TLC memory, with good endurance.
  19. Informative
    mariushm got a reaction from Aspirin77 in MSI B450M PRO-VDH MAX compatible with Ryzen 5 5600X?   
    Yes, that motherboard will most likely support the 5600x and the other processors in the series. 
     
    The VRM on that motherboard is good enough even for 8-12 core processors, if you don't overclock it. Even the 16 core processors will work on it, but I wouldn't recommend it without additional cooling on the VRM)
     
    Go on the website and see what bios updates the board has, there may be a bios update already with beta support for 5600x. 
     
    One thing you should check is if the board has BIOS Flashback feature, which would allow you to update bios without cpu installed.  If the mobo doesn't have this feature, you may need an older (compatible with installed bios version) processor in order to update bios to the desired version.
     
  20. Like
    mariushm got a reaction from SomeGuyNamedSean in My PSU came with 2 cpu 8 pin cables?   
    It's perfectly normal for a power supply to come with 2 CPU 8 pin
    Sometimes one of the cables can be split in 2  CPU 4 pin connectors, maybe that's why it says in smaller letters on one cable.
     
    This makes it possible to power a motherboard that has only one 4 pin CPU connector, or motherboards with one 8 pin and one 4 pin connector.
     
    Only ONE 8 pin is needed, if your motherboard has more than one 8 pin.  Anything else is optional, and only needed if you have very power hungry processors (think 12 or 16 core processors).
    If you don't like the extra cables, then your computer will be perfectly fine with just the 8 pin connector populated.
     
    A single 8 pin connector can give the processor over 300 watts. The 5600x will not consume more than 100-125 watts.  Extra power connectors are not needed when the first cable is barely used to half of its capabilities.
  21. Like
    mariushm got a reaction from toxsickcity in How does two dimms per channel work?   
    There's two memory channels ... in which you can have single rank , dual rank or even quad rank memory sticks (if the memory controller supports quad rank memory sticks) 
     
    Single Rank vs Dual Rank vs Quad Rank : 
     
    The ram stick is designed to receive or transmit 64 bits at a time to/from the processor. Well, 72 bits if you include ECC data, but for common desktop memory, we're talking 64 bits. 
    The tiny memory chips that are on the ram stick are designed to output either 4 bits or 8 bits at a time, these days chips that output 8 bits at a time are the most common.  
    So, in order to put 64 bits on the contacts of the ram stick, you need to have a multiple of 8 ram chips. If you use 4 bit ram chips, then you need 64/4 = 16 ram chips on the memory stick, so you'll have a single rank ram stick with 16 memory chips on it, probably 8 on each side. 
    If you use 8 bit ram chips, then you need at least 8 ram chips.  Manufacturer can then use either 8 or 16 ram chips to make the ram stick, so you could have a Single Rank ram stick with 8 ram chips on one side, or you could have a Dual Rank ram stick, with 2 groups of 8 ram chips.  (quad rank sticks would use 32 chips, 4x8 chips)
     
    The thing with Ranks is the processor can only work with a rank at a time, it can't read or write from both ranks at a time.
    But, it can do a think called interleaving.  
    Basically each time the processor wants something, it has to tell the ram chips : Guys , I want data at this position in your ram chips, get prepared to give this data to me. 
    Then, it takes a few nanoseconds for the ram chips in that rank to be ready to put the data on the ram stick pins, 64 bits at a time. 
    The processor can take advantage by this by sending a "get ready to give me data" to a rank right before it starts transferring data from another rank, for which it previously already sent the request to read or write data. 
    By the data the processor is done reading from that rank, the other rank spent those few nanoseconds preparing to send data and the processor can switch between ranks right away, without having to send the "get ready to transfer" command and wait nanoseconds doing nothing. 
     
    So dual rank ram sticks can give you a very slightly performance increase compared to single rank sticks, because with single rank sticks, every time a transfer is done, the processor has to send that "get ready to transfer" command and wait for the ram chips to being that transfer. With dual or quad rank, it can "chain" / interleave those "get ready calls" and skip on waiting those nanoseconds of time. 
     
    When you put two ram sticks on same channel, the memory controller basically treats each ram stick as if they're separate ranks ...  it knows the capacity of each ram stick on that channel and only enables the ram stick rank it wants to talk to. 
     
    -
     
    With dual channel, the processor takes advantage of having two sticks, one in each channel,  or four sticks , two in each channel, to transfer data to and from memory 128 bits at a time  (2 x 64 bit transfers).
     
    It still takes time (nonoseconds) for the "get ready to transfer command" to be "processed"  by the ram chips, but once that's done the processor gets 128 bits at a time, 64 bits from each memory channel. This way, it receives the desired data twice as fast. 
     
    Of course, you get most benefit when dealing with large amounts of data, because otherwise those nanoseconds of waiting time each time the processor has to send that "get ready to transfer" command can not be reduced. 
     
    For example, let's say it takes 10 ns to get ready and  4 ns to send  512 KB of data.  With dual channel, it would still take 10ns to get ready, but each channel will transfer half (256 KB) in 2ns, so overall it takes 10+2 = 12 ns instead of 14ns to transfer this data.  Dual channel in this case doesn't get you twice the speed, but it is a bit faster.
     
     
  22. Like
    mariushm got a reaction from johnyb98 in POS (point of sale) terminal   
    I would give it 192.168.1.101 or something further away from the other computers. 
    if you add or remove computers, if there's some power failure for example and computers which have DHCP enabled (auto get IP from router), they may turn on faster than the POS and "steal" that IP from the POS and then the POS will be stuck in a loop, unable to take the fixed IP address for itself. 
     
    DNS is only needed if you want the POS to have internet access. If so, use the IP of your ISP's DNS or use a public dns like Google's (8.8.8.8 / 8.8.4.4)  or Cloudflare (1.1.1.1) or OpenDNS  (208.67.222.222 or 208.67.220.220) .. these are just 3 examples. 
     
    Gateway ... leave empty (i'm not 100% sure) but you shouldn't need it.
     
    Only IP and subnet mask are needed for network communication.
     
    PS. Keep in mind that some routers which also have wireless, are configured by default to assign IPs to devices connected wireless  from a specific range, like for example starting from 192.168.1.201 onwards, or something like that. 
    basically, just keep a "buffer" between your fixed ip devices and ips that are assigned automatically to devices.
     
  23. Informative
    mariushm got a reaction from Eigenvektor in How does two dimms per channel work?   
    There's two memory channels ... in which you can have single rank , dual rank or even quad rank memory sticks (if the memory controller supports quad rank memory sticks) 
     
    Single Rank vs Dual Rank vs Quad Rank : 
     
    The ram stick is designed to receive or transmit 64 bits at a time to/from the processor. Well, 72 bits if you include ECC data, but for common desktop memory, we're talking 64 bits. 
    The tiny memory chips that are on the ram stick are designed to output either 4 bits or 8 bits at a time, these days chips that output 8 bits at a time are the most common.  
    So, in order to put 64 bits on the contacts of the ram stick, you need to have a multiple of 8 ram chips. If you use 4 bit ram chips, then you need 64/4 = 16 ram chips on the memory stick, so you'll have a single rank ram stick with 16 memory chips on it, probably 8 on each side. 
    If you use 8 bit ram chips, then you need at least 8 ram chips.  Manufacturer can then use either 8 or 16 ram chips to make the ram stick, so you could have a Single Rank ram stick with 8 ram chips on one side, or you could have a Dual Rank ram stick, with 2 groups of 8 ram chips.  (quad rank sticks would use 32 chips, 4x8 chips)
     
    The thing with Ranks is the processor can only work with a rank at a time, it can't read or write from both ranks at a time.
    But, it can do a think called interleaving.  
    Basically each time the processor wants something, it has to tell the ram chips : Guys , I want data at this position in your ram chips, get prepared to give this data to me. 
    Then, it takes a few nanoseconds for the ram chips in that rank to be ready to put the data on the ram stick pins, 64 bits at a time. 
    The processor can take advantage by this by sending a "get ready to give me data" to a rank right before it starts transferring data from another rank, for which it previously already sent the request to read or write data. 
    By the data the processor is done reading from that rank, the other rank spent those few nanoseconds preparing to send data and the processor can switch between ranks right away, without having to send the "get ready to transfer" command and wait nanoseconds doing nothing. 
     
    So dual rank ram sticks can give you a very slightly performance increase compared to single rank sticks, because with single rank sticks, every time a transfer is done, the processor has to send that "get ready to transfer" command and wait for the ram chips to being that transfer. With dual or quad rank, it can "chain" / interleave those "get ready calls" and skip on waiting those nanoseconds of time. 
     
    When you put two ram sticks on same channel, the memory controller basically treats each ram stick as if they're separate ranks ...  it knows the capacity of each ram stick on that channel and only enables the ram stick rank it wants to talk to. 
     
    -
     
    With dual channel, the processor takes advantage of having two sticks, one in each channel,  or four sticks , two in each channel, to transfer data to and from memory 128 bits at a time  (2 x 64 bit transfers).
     
    It still takes time (nonoseconds) for the "get ready to transfer command" to be "processed"  by the ram chips, but once that's done the processor gets 128 bits at a time, 64 bits from each memory channel. This way, it receives the desired data twice as fast. 
     
    Of course, you get most benefit when dealing with large amounts of data, because otherwise those nanoseconds of waiting time each time the processor has to send that "get ready to transfer" command can not be reduced. 
     
    For example, let's say it takes 10 ns to get ready and  4 ns to send  512 KB of data.  With dual channel, it would still take 10ns to get ready, but each channel will transfer half (256 KB) in 2ns, so overall it takes 10+2 = 12 ns instead of 14ns to transfer this data.  Dual channel in this case doesn't get you twice the speed, but it is a bit faster.
     
     
  24. Like
    mariushm got a reaction from SomeGuyNamedSean in Really dumb thermalpaste question   
    How much time is between application and first start up doesn't matter. Won't make any difference.
     
    The paste will be trapped in the impurities and denivelations in the two metal surfaces and help transfer the heat between the two metals.  The amount of evaporation going on on the edges of the heatsinks is extremely small, practically doesn't matter.
     
    A rice or a pea size is a bit outdated, you can put a bit more.  Too little is bad.  A bit more than ideal won't make any meaningful difference, it's like maybe cpu being 0.1..0.3c hotter than ideal case. A LOT is a bit worse... but you'd really have to go overboard with paste to be A LOT.
     
    You can apply a bit more, let's say about as much as you see people in toothpaste ads apply toothpaste on a toothbrush. 
     
    edit: don't spread the thermal paste yourself  as in don't use a ruler/card/finger to paint the cpu surface with thermal paste ... put a big blob in the middle, or make an X on the surface of the cpu, basically two lines of paste crossing each other. When you put down the cooler and gently press it down, it will flatten the paste and push it towards the edges so you won't have any trapped air bubbles between cpu and heatsink. 
     
  25. Like
    mariushm got a reaction from mihaivoicu55 in (UK) Hello, it is my sister's birthday in 5 days and we were looking to get her a laptop at no more than £500   
    It's a bit tough. 
    At that price you'll have a lot of laptop models that will not be upgradeable ... for example will come with ram soldered, and  no extra ram slots to go from 8 GB to 16 GB or from 16 GB to 32 GB. Also, you won't get the standard RJ45 (ethernet) port, only wireless connections. 
     
    If you can stretch it a bit, here an ok model : 
     
    551 GBP, in stock : https://www.amazon.co.uk/HP-15s-fq1003na-15-6-Inch-Laptop/dp/B08CDY1KJZ/
     
    Has one of the latest Intel processors, fast 512 GB storage, decent screen and camera, but 8 GB and wireless only.  Fine for browsing and minecraft and other such games and programming.  Note the shipping though, fastest they estimate is 26th of January.
     
    If you can somehow order them (they seem to be temporarily out of stock) these Asus VivoBook laptops seem to have decent specifications for the money : 
     
    549 - 449 GBP
     
    https://www.amazon.co.uk/ASUS-VivoBook-Laptop-Windows-reversible/dp/B089Q2PWN8/
    https://www.amazon.co.uk/ASUS-VivoBook-X512DA-Laptop-Windows/dp/B0845XZ5FQ/
    https://www.amazon.co.uk/ASUS-VivoBook-X512DA-R5-3500-Windows/dp/B07SHT8C3D/
     
     
     
×