• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Recent content by LukeTbk

  1. L

    RX 9050 is a bit of a turd

    yes that what i mean by same gpu die.... 50% for such a small for a gpu chip on such an old and mature node seem really aggressive, specially with how precious spending 8GB of semi-descent gddr6 is at the moment.
  2. L

    RX 9050 is a bit of a turd

    oem in mind move can be hard to understand of course, but this one feel like I am missing something. It is using the same 18Gbps memory according to what I can find than a 9060 would, it is using the same GPU die, why going with such an aggressive core count cut (50%), at under 200mm on a very...
  3. L

    DRAM Pricing Jumps 50%, Only 70% of Orders Getting Filled

    the amount of those that goes into a 16GB of memory (glass, Tin, Tungsten and others) is minisimal and the level of purety/rafining/processing on them before they can be used in memory (and how costly to process to make memory), make it that memory cannot ever impact raw reserve. The most...
  4. L

    DRAM Pricing Jumps 50%, Only 70% of Orders Getting Filled

    Silicon is one of the most abudent thing there is (28% of earth crust....) and we make terabyte of memory on a small 12inch wafer, I think this was a joke. the ratio of complexity / cost of turning the raw material into memory relative to how extremelly common they are make it not something...
  5. L

    Zen 7 Rumours / AMD "Zen 6" ISA to Bring AVX512 FP16, VNNI INT8, and More

    Maybe in the near future, but right now Epyc is incredible. Look at the price tag per mm (of very high yield small chiplet) of silicon they can sales Epyc for and they sales the very expensive motherboard controller. CDNA-helios were still at nearly a lost effort not so long ago I think and...
  6. L

    Anthropic’s Mythos AI used social engineering to target real people

    The meme does not really work here, Anthropic was not in any way benefiting from the agents action (if they would have worked, which did not in this case), of course if an AI agent steal from a banks they will press charge. No damage was done by an AI agents in the example from the bbc article.
  7. L

    Anthropic’s Mythos AI used social engineering to target real people

    This they are heavily train to reward being cheap to run as much as possible. Using social engineering hacking will often be the lowest effort / cheapest way to do the type of task (when that kind of story occur, it seem to be the driving force) and they do it for the same reason human attacker...
  8. L

    AI LLM hardware discussion thread

    True enough, with that many * maybe that use case specific, better just: https://aws.amazon.com/ec2/instance-types/mac/ If you can, actually try them before pulling the trigger.
  9. L

    AI LLM hardware discussion thread

    For going 256GB, i would look Apple m3 ultra 256gb price vs 2x DGX spark... The memory bandwith will be quite higher on apple here, ~800 GB/s, no FP4 support of course, Spark get more interesting if you know in advance you will use only nvpf4 models. and apparently now support have been added...
  10. L

    AI LLM hardware discussion thread

    sparks are nice for DEV in the cuda env, but lot of money for that very limited level of bandwith if you only interested in running things, depend quite a bit what you want to do. One general tips, would be to try renting ont he cloud (5060, 3090, etc... are the type you can) to run your exact...
  11. L

    AI LLM hardware discussion thread

    depending on what you do-need size wise, used single 3090 or M2 era 24GB mac mini would be alternative... a single 3090 still have more memory bandwith than 2 modern but 128 bits gpu, blackwell will offer "native" modern FP4 support and other bonus, but can be a bit more complicated setup...
  12. L

    How AI will fail

    they cannot break an particulary contract they signed with a particular employee is not some large rule about firing employee due to AI, the precendant set here is thet cannot use simple LLM able to do what a worker could to use the force majeur-catasthropic unpredicted event close those...
  13. L

    Are all software platforms going to cloud?

    I really doubt they were digitising that data for 70 years;), would it have been scan of mediaval painting it is not that relevant , what matter to judge the quality of their IT decision making is what was the situation when they decided to put their data only on a single small playerr cloud...
  14. L

    Are all software platforms going to cloud?

    Has we get old, very recently can mean different things (Sometime I feel 2008 was...), but it has been for at least a decade and that contract with OSS is only since 2019. In 2014, 64 TB, using 4TB sweet spot would have been just ~$3000, with a nice pro-issh NAS still talking around $5000 for...
  15. L

    Are all software platforms going to cloud?

    50 terabyte is such a ridiculously small amount of data, that would cost virtually zero to have a local version, it would be strange to not have it (3 regular USB harddrive can hold all of that....), having 3 copy in 3 different part of the building would still be really cheap. We are talking...
See more
Back
Top