Jump to content

LAwLz

Member
  • Content count

    11,506
  • Joined

  • Last visited


Reputation Activity

  1. Informative
    LAwLz got a reaction from tjcater for a status update, My mother's ISP ( Bahnhof, the same ones that reported themselves to the police for r   
    My mother's ISP (Bahnhof, the same ones that reported themselves to the police for refusing to spy on customer) has just launched 10Gbps home Internet connections.
    If you sign up for 2 years it costs 37 USD a month for the first 6 months, and after that it's 60 USD a month.
     
    60 USD a month for 10Gbps Internet connection.
     
    They have also lowered the cost of their 1Gbps package to 50 dollars a month.
     
    They are also offering 40Gbps and 100Gbps Internet connections to enterprise customers.
     
    And no, I did not confuse the G with an M.
     
    https://www.bahnhof.se/press/press-releases/2018/02/21/bahnhof-lanserar-10-000-mbit-s-till-privatpersoner-for-298-kr-man
    https://www.bahnhof.se/villafiber/
  2. Informative
    LAwLz got a reaction from tjcater for a status update, My mother's ISP ( Bahnhof, the same ones that reported themselves to the police for r   
    My mother's ISP (Bahnhof, the same ones that reported themselves to the police for refusing to spy on customer) has just launched 10Gbps home Internet connections.
    If you sign up for 2 years it costs 37 USD a month for the first 6 months, and after that it's 60 USD a month.
     
    60 USD a month for 10Gbps Internet connection.
     
    They have also lowered the cost of their 1Gbps package to 50 dollars a month.
     
    They are also offering 40Gbps and 100Gbps Internet connections to enterprise customers.
     
    And no, I did not confuse the G with an M.
     
    https://www.bahnhof.se/press/press-releases/2018/02/21/bahnhof-lanserar-10-000-mbit-s-till-privatpersoner-for-298-kr-man
    https://www.bahnhof.se/villafiber/
  3. Informative
    LAwLz got a reaction from tjcater for a status update, My mother's ISP ( Bahnhof, the same ones that reported themselves to the police for r   
    My mother's ISP (Bahnhof, the same ones that reported themselves to the police for refusing to spy on customer) has just launched 10Gbps home Internet connections.
    If you sign up for 2 years it costs 37 USD a month for the first 6 months, and after that it's 60 USD a month.
     
    60 USD a month for 10Gbps Internet connection.
     
    They have also lowered the cost of their 1Gbps package to 50 dollars a month.
     
    They are also offering 40Gbps and 100Gbps Internet connections to enterprise customers.
     
    And no, I did not confuse the G with an M.
     
    https://www.bahnhof.se/press/press-releases/2018/02/21/bahnhof-lanserar-10-000-mbit-s-till-privatpersoner-for-298-kr-man
    https://www.bahnhof.se/villafiber/
  4. Informative
    LAwLz got a reaction from Cinnabar Sonar for a status update, Just a heads up, the Spectre protection flag in Visual Studio will only provide a ver   
    Just a heads up, the Spectre protection flag in Visual Studio will only provide a very weak and narrow protection against Spectre.
     
    Good article about why can be found here:
    Microsoft’s compiler-level Spectre fix shows how hard this problem will be to solve
  5. Informative
    LAwLz got a reaction from Cinnabar Sonar for a status update, Just a heads up, the Spectre protection flag in Visual Studio will only provide a ver   
    Just a heads up, the Spectre protection flag in Visual Studio will only provide a very weak and narrow protection against Spectre.
     
    Good article about why can be found here:
    Microsoft’s compiler-level Spectre fix shows how hard this problem will be to solve
  6. Informative
    LAwLz got a reaction from Cinnabar Sonar for a status update, Just a heads up, the Spectre protection flag in Visual Studio will only provide a ver   
    Just a heads up, the Spectre protection flag in Visual Studio will only provide a very weak and narrow protection against Spectre.
     
    Good article about why can be found here:
    Microsoft’s compiler-level Spectre fix shows how hard this problem will be to solve
  7. Like
    LAwLz got a reaction from Yoshcko for a status update, I love watching videos analyzing things, especially things you have never really thou   
    I love watching videos analyzing things, especially things you have never really thought of before, and this one is a perfect example of that.
    It's about sharing headphones and why it's romantic.
    It's short (3,5 minutes), really well made, and makes well thought out comments about society. Highly recommend checking out.
     
  8. Like
    LAwLz got a reaction from Yoshcko for a status update, I love watching videos analyzing things, especially things you have never really thou   
    I love watching videos analyzing things, especially things you have never really thought of before, and this one is a perfect example of that.
    It's about sharing headphones and why it's romantic.
    It's short (3,5 minutes), really well made, and makes well thought out comments about society. Highly recommend checking out.
     
  9. Like
    LAwLz got a reaction from Yoshcko for a status update, I love watching videos analyzing things, especially things you have never really thou   
    I love watching videos analyzing things, especially things you have never really thought of before, and this one is a perfect example of that.
    It's about sharing headphones and why it's romantic.
    It's short (3,5 minutes), really well made, and makes well thought out comments about society. Highly recommend checking out.
     
  10. Agree
    LAwLz got a reaction from Yoshcko for a status update, I wish they taught cinematography in schools. When I went to school, we were taught w   
    I wish they taught cinematography in schools.
    When I went to school, we were taught woodworking, sewing, music and drawing. I think one of these could be (optionally) replaced with cinematography.
     
    I'm not talking about "let's watch a movie and discuss what the message was". I'm talking about how the scene was composed, how the movie was edited, tempo, structure and other things like that.
    In our drawing classes we learned about perspective, color mixing, proportions and other things. I think it's about time a similar subject appears for movies, because it seems like most people severely lacks in this area. I've certainly realized how ignorant I am about it after listening and talking to people who have been to film school.
     
    It also reminded me about the comment made by Luke during one of the WAN shows. He said that he loves going to the cinema with Brandon because he asks questions afterward that made Luke think about how the movie was made, questions like "why do you think you liked that scene in particular?".
     
    I think teaching these things could help people appreciate movies more, and maybe inspire people to make their own movies as well.
  11. Agree
    LAwLz got a reaction from Yoshcko for a status update, I wish they taught cinematography in schools. When I went to school, we were taught w   
    I wish they taught cinematography in schools.
    When I went to school, we were taught woodworking, sewing, music and drawing. I think one of these could be (optionally) replaced with cinematography.
     
    I'm not talking about "let's watch a movie and discuss what the message was". I'm talking about how the scene was composed, how the movie was edited, tempo, structure and other things like that.
    In our drawing classes we learned about perspective, color mixing, proportions and other things. I think it's about time a similar subject appears for movies, because it seems like most people severely lacks in this area. I've certainly realized how ignorant I am about it after listening and talking to people who have been to film school.
     
    It also reminded me about the comment made by Luke during one of the WAN shows. He said that he loves going to the cinema with Brandon because he asks questions afterward that made Luke think about how the movie was made, questions like "why do you think you liked that scene in particular?".
     
    I think teaching these things could help people appreciate movies more, and maybe inspire people to make their own movies as well.
  12. Agree
    LAwLz got a reaction from Yoshcko for a status update, I wish they taught cinematography in schools. When I went to school, we were taught w   
    I wish they taught cinematography in schools.
    When I went to school, we were taught woodworking, sewing, music and drawing. I think one of these could be (optionally) replaced with cinematography.
     
    I'm not talking about "let's watch a movie and discuss what the message was". I'm talking about how the scene was composed, how the movie was edited, tempo, structure and other things like that.
    In our drawing classes we learned about perspective, color mixing, proportions and other things. I think it's about time a similar subject appears for movies, because it seems like most people severely lacks in this area. I've certainly realized how ignorant I am about it after listening and talking to people who have been to film school.
     
    It also reminded me about the comment made by Luke during one of the WAN shows. He said that he loves going to the cinema with Brandon because he asks questions afterward that made Luke think about how the movie was made, questions like "why do you think you liked that scene in particular?".
     
    I think teaching these things could help people appreciate movies more, and maybe inspire people to make their own movies as well.
  13. Like
    LAwLz got a reaction from wkdpaul for a status update, Guess it's time to mine Ethereum. I'l just do it on my single 1060 during the nights   
    Guess it's time to mine Ethereum. I'l just do it on my single 1060 during the nights but hopefully it will add up to a little bit. At the very least I won't feel like an idiot if it takes off like bitcoin did. 
  14. Like
    LAwLz got a reaction from wkdpaul for a status update, Guess it's time to mine Ethereum. I'l just do it on my single 1060 during the nights   
    Guess it's time to mine Ethereum. I'l just do it on my single 1060 during the nights but hopefully it will add up to a little bit. At the very least I won't feel like an idiot if it takes off like bitcoin did. 
  15. Like
    LAwLz got a reaction from wkdpaul for a status update, Guess it's time to mine Ethereum. I'l just do it on my single 1060 during the nights   
    Guess it's time to mine Ethereum. I'l just do it on my single 1060 during the nights but hopefully it will add up to a little bit. At the very least I won't feel like an idiot if it takes off like bitcoin did. 
  16. Like
    LAwLz got a reaction from wkdpaul for a status update, Guess it's time to mine Ethereum. I'l just do it on my single 1060 during the nights   
    Guess it's time to mine Ethereum. I'l just do it on my single 1060 during the nights but hopefully it will add up to a little bit. At the very least I won't feel like an idiot if it takes off like bitcoin did. 
  17. Like
    LAwLz got a reaction from PCGuy_5960 for a status update, For those wondering how Meltdown and Spectre works, here is a very simplified explana   
    For those wondering how Meltdown and Spectre works, here is a very simplified explanation (to my understanding of it at least, which might be wrong):
    Let's say we have two arrays of data.
    A = X[13]
    B = Y[A]
     
    As you can see, in order to know what B is, it needs to fetch data from A.
    On modern processors, B can be loaded before it has checked if A is safe to access. This is where the "speculation" of the processor comes in because it does not actually know if calculating B is acceptable or not. It might discover some issue and need to throw all that work away, or it might have saved some time by doing calculations earlier than it otherwise would have.
     
    What will happen is that A is loaded, then Y[A] is calculated and stored as B.
    What happens when the processor discovers that A is protected memory, such as kernel memory?
    The processor will just roll things back and pretend like it never accessed Y[A], but the problem is that B was loaded into the cache. By doing some clever tricks (such as measuring latencies), it is possible to determine if the bit used for the prediction in A (which is protected memory) was a 1 or a 0.
    Put this into a loop and you can read the entire A, one bit at a time.
     
    The other exploits uses similar techniques but instead of using things like speculative loads between security rings (A and B in this example were in different security rings), it instead uses the speculation of things like branch prediction.
     
    The general idea of all three exploits are the same. Create a scenario where the prediction/speculation of processors leaves clues of what data was accessed before it realized the predictions/speculations were invalid.
    Some methods of finding "clues" are easier to execute, and the clues are more obvious, but the concepts are the same.
     
    What's a bit worrying is that three different methods of "finding clues" were published, but this is essentially a brand new category of attacks. There might be hundreds of ways of abusing this concept, and verifying that a design is safe against these types of attacks is extremely difficult.
  18. Like
    LAwLz got a reaction from PCGuy_5960 for a status update, For those wondering how Meltdown and Spectre works, here is a very simplified explana   
    For those wondering how Meltdown and Spectre works, here is a very simplified explanation (to my understanding of it at least, which might be wrong):
    Let's say we have two arrays of data.
    A = X[13]
    B = Y[A]
     
    As you can see, in order to know what B is, it needs to fetch data from A.
    On modern processors, B can be loaded before it has checked if A is safe to access. This is where the "speculation" of the processor comes in because it does not actually know if calculating B is acceptable or not. It might discover some issue and need to throw all that work away, or it might have saved some time by doing calculations earlier than it otherwise would have.
     
    What will happen is that A is loaded, then Y[A] is calculated and stored as B.
    What happens when the processor discovers that A is protected memory, such as kernel memory?
    The processor will just roll things back and pretend like it never accessed Y[A], but the problem is that B was loaded into the cache. By doing some clever tricks (such as measuring latencies), it is possible to determine if the bit used for the prediction in A (which is protected memory) was a 1 or a 0.
    Put this into a loop and you can read the entire A, one bit at a time.
     
    The other exploits uses similar techniques but instead of using things like speculative loads between security rings (A and B in this example were in different security rings), it instead uses the speculation of things like branch prediction.
     
    The general idea of all three exploits are the same. Create a scenario where the prediction/speculation of processors leaves clues of what data was accessed before it realized the predictions/speculations were invalid.
    Some methods of finding "clues" are easier to execute, and the clues are more obvious, but the concepts are the same.
     
    What's a bit worrying is that three different methods of "finding clues" were published, but this is essentially a brand new category of attacks. There might be hundreds of ways of abusing this concept, and verifying that a design is safe against these types of attacks is extremely difficult.
  19. Like
    LAwLz got a reaction from PCGuy_5960 for a status update, For those wondering how Meltdown and Spectre works, here is a very simplified explana   
    For those wondering how Meltdown and Spectre works, here is a very simplified explanation (to my understanding of it at least, which might be wrong):
    Let's say we have two arrays of data.
    A = X[13]
    B = Y[A]
     
    As you can see, in order to know what B is, it needs to fetch data from A.
    On modern processors, B can be loaded before it has checked if A is safe to access. This is where the "speculation" of the processor comes in because it does not actually know if calculating B is acceptable or not. It might discover some issue and need to throw all that work away, or it might have saved some time by doing calculations earlier than it otherwise would have.
     
    What will happen is that A is loaded, then Y[A] is calculated and stored as B.
    What happens when the processor discovers that A is protected memory, such as kernel memory?
    The processor will just roll things back and pretend like it never accessed Y[A], but the problem is that B was loaded into the cache. By doing some clever tricks (such as measuring latencies), it is possible to determine if the bit used for the prediction in A (which is protected memory) was a 1 or a 0.
    Put this into a loop and you can read the entire A, one bit at a time.
     
    The other exploits uses similar techniques but instead of using things like speculative loads between security rings (A and B in this example were in different security rings), it instead uses the speculation of things like branch prediction.
     
    The general idea of all three exploits are the same. Create a scenario where the prediction/speculation of processors leaves clues of what data was accessed before it realized the predictions/speculations were invalid.
    Some methods of finding "clues" are easier to execute, and the clues are more obvious, but the concepts are the same.
     
    What's a bit worrying is that three different methods of "finding clues" were published, but this is essentially a brand new category of attacks. There might be hundreds of ways of abusing this concept, and verifying that a design is safe against these types of attacks is extremely difficult.
  20. Like
    LAwLz got a reaction from PCGuy_5960 for a status update, For those wondering how Meltdown and Spectre works, here is a very simplified explana   
    For those wondering how Meltdown and Spectre works, here is a very simplified explanation (to my understanding of it at least, which might be wrong):
    Let's say we have two arrays of data.
    A = X[13]
    B = Y[A]
     
    As you can see, in order to know what B is, it needs to fetch data from A.
    On modern processors, B can be loaded before it has checked if A is safe to access. This is where the "speculation" of the processor comes in because it does not actually know if calculating B is acceptable or not. It might discover some issue and need to throw all that work away, or it might have saved some time by doing calculations earlier than it otherwise would have.
     
    What will happen is that A is loaded, then Y[A] is calculated and stored as B.
    What happens when the processor discovers that A is protected memory, such as kernel memory?
    The processor will just roll things back and pretend like it never accessed Y[A], but the problem is that B was loaded into the cache. By doing some clever tricks (such as measuring latencies), it is possible to determine if the bit used for the prediction in A (which is protected memory) was a 1 or a 0.
    Put this into a loop and you can read the entire A, one bit at a time.
     
    The other exploits uses similar techniques but instead of using things like speculative loads between security rings (A and B in this example were in different security rings), it instead uses the speculation of things like branch prediction.
     
    The general idea of all three exploits are the same. Create a scenario where the prediction/speculation of processors leaves clues of what data was accessed before it realized the predictions/speculations were invalid.
    Some methods of finding "clues" are easier to execute, and the clues are more obvious, but the concepts are the same.
     
    What's a bit worrying is that three different methods of "finding clues" were published, but this is essentially a brand new category of attacks. There might be hundreds of ways of abusing this concept, and verifying that a design is safe against these types of attacks is extremely difficult.
  21. Like
    LAwLz got a reaction from PCGuy_5960 for a status update, For those wondering how Meltdown and Spectre works, here is a very simplified explana   
    For those wondering how Meltdown and Spectre works, here is a very simplified explanation (to my understanding of it at least, which might be wrong):
    Let's say we have two arrays of data.
    A = X[13]
    B = Y[A]
     
    As you can see, in order to know what B is, it needs to fetch data from A.
    On modern processors, B can be loaded before it has checked if A is safe to access. This is where the "speculation" of the processor comes in because it does not actually know if calculating B is acceptable or not. It might discover some issue and need to throw all that work away, or it might have saved some time by doing calculations earlier than it otherwise would have.
     
    What will happen is that A is loaded, then Y[A] is calculated and stored as B.
    What happens when the processor discovers that A is protected memory, such as kernel memory?
    The processor will just roll things back and pretend like it never accessed Y[A], but the problem is that B was loaded into the cache. By doing some clever tricks (such as measuring latencies), it is possible to determine if the bit used for the prediction in A (which is protected memory) was a 1 or a 0.
    Put this into a loop and you can read the entire A, one bit at a time.
     
    The other exploits uses similar techniques but instead of using things like speculative loads between security rings (A and B in this example were in different security rings), it instead uses the speculation of things like branch prediction.
     
    The general idea of all three exploits are the same. Create a scenario where the prediction/speculation of processors leaves clues of what data was accessed before it realized the predictions/speculations were invalid.
    Some methods of finding "clues" are easier to execute, and the clues are more obvious, but the concepts are the same.
     
    What's a bit worrying is that three different methods of "finding clues" were published, but this is essentially a brand new category of attacks. There might be hundreds of ways of abusing this concept, and verifying that a design is safe against these types of attacks is extremely difficult.
  22. Like
    LAwLz got a reaction from PCGuy_5960 for a status update, For those wondering how Meltdown and Spectre works, here is a very simplified explana   
    For those wondering how Meltdown and Spectre works, here is a very simplified explanation (to my understanding of it at least, which might be wrong):
    Let's say we have two arrays of data.
    A = X[13]
    B = Y[A]
     
    As you can see, in order to know what B is, it needs to fetch data from A.
    On modern processors, B can be loaded before it has checked if A is safe to access. This is where the "speculation" of the processor comes in because it does not actually know if calculating B is acceptable or not. It might discover some issue and need to throw all that work away, or it might have saved some time by doing calculations earlier than it otherwise would have.
     
    What will happen is that A is loaded, then Y[A] is calculated and stored as B.
    What happens when the processor discovers that A is protected memory, such as kernel memory?
    The processor will just roll things back and pretend like it never accessed Y[A], but the problem is that B was loaded into the cache. By doing some clever tricks (such as measuring latencies), it is possible to determine if the bit used for the prediction in A (which is protected memory) was a 1 or a 0.
    Put this into a loop and you can read the entire A, one bit at a time.
     
    The other exploits uses similar techniques but instead of using things like speculative loads between security rings (A and B in this example were in different security rings), it instead uses the speculation of things like branch prediction.
     
    The general idea of all three exploits are the same. Create a scenario where the prediction/speculation of processors leaves clues of what data was accessed before it realized the predictions/speculations were invalid.
    Some methods of finding "clues" are easier to execute, and the clues are more obvious, but the concepts are the same.
     
    What's a bit worrying is that three different methods of "finding clues" were published, but this is essentially a brand new category of attacks. There might be hundreds of ways of abusing this concept, and verifying that a design is safe against these types of attacks is extremely difficult.
  23. Like
    LAwLz reacted to Sauron for a status update, Happy new year's eve (or new year depending on where you live)   
    Happy new year's eve (or new year depending on where you live)
  24. Informative
    LAwLz got a reaction from Xiauj for a status update, I feel really disconnected from other star wars fans right now. I did not like episod   
    I feel really disconnected from other star wars fans right now.
    I did not like episode 8 but it is getting really high scores and my friends are saying I got bad taste for not liking it. 
     
    I would give it a 6/10 while some of them say it was the best star wars movie ever. 
  25. Informative
    LAwLz got a reaction from Xiauj for a status update, I feel really disconnected from other star wars fans right now. I did not like episod   
    I feel really disconnected from other star wars fans right now.
    I did not like episode 8 but it is getting really high scores and my friends are saying I got bad taste for not liking it. 
     
    I would give it a 6/10 while some of them say it was the best star wars movie ever. 
×