Support NeoGAF

[TheJollyCorner]

Just wait for Rockstar to sort this out, seriously. This is a much more narrative based game than people originally anticipated I think.

BenjaminBirdie hath spoken... and thus BenjaminBirdie's will be done! I shall act on this... for I have nothing left to try.


...don't cock this up any further, Rock*. I feel like the cripple kid sitting on the sidelines while the other kids are playing football.

I WANT TO GO IN! TAKE THE LEG BRACES OFF ME AND HAND ME A JOCK STRAP!

 

[JasoNsider]

wow, you do not deserve to play this game.

Apparently neither do some people who have been waiting since the game was announced :lol Seriously now, spare me the old line. People know QA is an issue we absolutely cannot bullshit around with in the this industry, and I would give equal amounts of shit to any company that releases unstable products.

 

[jetjevons]

Great news. The disc that got passed the original cut-scene on my home PS3 just locked up and crashed during the

Spoiler

michelle date

mission. Yay!

 

[GeoramA]

Just played for 5 hours straight. No freezing this time.

 

[Baker]

I shall act on this... for I have nothing left to try.

Same. They just better say something/anything tomorrow or I'm returning the game.

 

[Killthee]

I was wondering that too. I really don't like the sound it makes when it shuts down after the lock up.

The Warhawk beta made my system (60GB Launch; I upgraded the drive a couple of months after the beta ended) crash +50 times (probably +75) and my system hasn't had any problems or issues with GTA4.

 

[BenjaminBirdie]

BenjaminBirdie hath spoken... and thus BenjaminBirdie's will be done! I shall act on this... for I have nothing left to try.

:lol

Trust me, though. It's such a cohesive and great narrative experience, you'd be doing yourself a disservice.

 

[mildewproduction]

I'm curious how many multi-console owners are selling the PS3 version and going with 360?

 

[jetjevons]

I'm curious how many multi-console owners are selling the PS3 version and going with 360?

Maybe this is what MS was really buying with that 50 million. The curse of Jack Thompson on the PS3 sku!

 

[Speevy]

no problems yet, 60 gb

 

[Dante]

I'm curious how many multi-console owners are selling the PS3 version and going with 360?

Are we reading the same thread? Because it doesn't sound like any system is a completely safe bet to not have lock ups at this point.

 

[dreamer3kx]

Just played for 5 hours straight. No freezing this time.

man how can you guys play for so long, I have a hard time playing more then a couple of hours.

 

[TheDarkKnight774]

froze in the early intro cutscenes. Launch PS3 60gb

but after that i played 1:30 hours and no issues.

 

[LowParry]

still no word from R* on a fix or something? Just "we are aware of the problem"?

 

[NinjaCodah]

Apparently neither do some people who have been waiting since the game was announced :lol Seriously now, spare me the old line. People know QA is an issue we absolutely cannot bullshit around with in the this industry, and I would give equal amounts of shit to any company that releases unstable products.

If it would be that unstable, EVERYBODY would have problems with it.

oh btw, I had a crash once in MGS3, will you dare take a dump on that game as well now, or on Kojima Production?

 

[ComputerNerd]

Strange how fast you guys want a fix. Ya know, sometimes it takes a while to find a bug.

There may be multiple bugs.

 

[LowParry]

Strange how fast you guys want a fix. Ya know, sometimes it takes a while to find a bug.

There may be multiple bugs.

We all want the best experience a game can give us. If its got problems, a quick fix should be the ideal thing to come. Especially a game that's this big. That's just me though.

 

[ggnoobIGN]

Strange how fast you guys want a fix. Ya know, sometimes it takes a while to find a bug.

There may be multiple bugs.

Want a fix? We wanted a game that works to begin with. We have a right to be impatient and very upset with Rockstar because they have handled this terribly, no denying that.

 

[NinjaCodah]

Strange how fast you guys want a fix. Ya know, sometimes it takes a while to find a bug.

There may be multiple bugs.

No kidding, the game is not even out 24hours and people are already worried that Rockstar hasn't addressed the issue yet.

 

[teh_pwn]

Is this game using multiple threads on all of Cell's 8 cores? The description of it happening rarely on some 60 GB models make me think it's a software race condition.

We all want the best experience a game can give us. If its got problems, a quick fix should be the ideal thing to come. Especially a game that's this big. That's just me though.

Unless R* has some way of getting bug reports, this could take a long time to track down. The code behind something like that isn't simple.

 

[vilmer_]

Just an update. Played for a good solid 3 hours non-stop, no total lock-ups at all. The only problem I ran into was a about a 3 second hitch when the subway train was coming to pick me up. 20gb launch.

 

[ggnoobIGN]

No kidding, the game is not even out 24hours and people are already worried that Rockstar hasn't addressed the issue yet.

It is absolutely mind boggling that people are defending Rockstar with regards to this situation in any way, shape, or form.

 

[LowParry]

Well. I'm packing the game up for now till they get it fixed. Will be a ball buster but I'll have some other games to finish up to tide me over. And a few other gems to play.

 

[ComputerNerd]

It is absolutely mind boggling that people are defending Rockstar with regards to this situation in any way, shape, or form.

I'm not denying that this situation should have been known about and fixed prior to release. But at this point, you'll have to wait a few days probably for a clear response from Rockstar. They need to track down the problem.

Don't get a heart attack, and wait a few days.

 

[teh_pwn]

It is absolutely mind boggling that people are defending Rockstar with regards to this situation in any way, shape, or form.

Defending? I'm being realistic. Complex multi-threading tasks that very rarely cause crashes on particular models of variations of hardware is not something you can quick fix. Unless Rockstar has some way of getting a bug report that gives them a stack crawl and a line number, this could take a long ass time to fix if it is a race condition.

A race condition is when you have a condition in multi-threading code such that two threads operate out of order depending on variations of conditions or hardware. It's when something that is supposed to rigidly be defined in data flow operates very rarely out of order. It's extremely hard to find because it requires you to start by looking at the entire source. You've got to have something to narrow down something massive like this. It's possible that a race condition could work in testing, but with a slight variation in hardware architecture (revision) that something happen out of order too quickly.

 

[ScrabbleBanshee]

In other news, I'm bringing my Toyota in for a recall tomorrow afternoon. The world is truly going up in flames around us.

 

[ggnoobIGN]

I'm not denying that this situation should have been known about and fixed prior to release. But at this point, you'll have to wait a few days probably for a clear response from Rockstar. They need to track down the problem.

Don't get a heart attack, and wait a few days.

No thanks. I, like many others, would rather not let RockStar off the hook for such a disaster. The more we all complain, the more pressure is on them to actually do something about this. Hopefully this gets into the news(It already has to an extent), and gives RockStar a blackeye for such a massive screw up.

This marks 2 games in a month that have shipped with major bugs(I don't have Bully, but that is what I am told). Simply inexcusable.

 

[ggnoobIGN]

Defending? I'm being realistic. Complex multi-threading tasks that very rarely cause crashes on particular models of variations of hardware is not something you can quick fix. Unless Rockstar has some way of getting a bug report that gives them a stack crawl and a line number, this could take a long ass time to fix if it is a race condition.

A race condition is when you have a condition in multi-threading code such that two shared resources occasionally don't happen in order. It's possible that a race condition could work in testing, but with a slight variation in architecture (revision) that something happen out of order too quickly.

My post wasn't referenced to you. I have liked the unqiue info you have provided in regards to the situation(although hopefully it isn't true and something can get done faster).

 

[teh_pwn]

My post wasn't referenced to you. I have liked the unqiue info you have provided in regards to the situation(although hopefully it isn't true and something can get done faster).

I edited my post. It doesn't have to be caused by a shared resource.

 

[ScrabbleBanshee]

Defending? I'm being realistic. Complex multi-threading tasks that very rarely cause crashes on particular models of variations of hardware is not something you can quick fix. Unless Rockstar has some way of getting a bug report that gives them a stack crawl and a line number, this could take a long ass time to fix if it is a race condition.

A race condition is when you have a condition in multi-threading code such that two shared resources occasionally don't happen in order. It's possible that a race condition could work in testing, but with a slight variation in architecture (revision) that something happen out of order too quickly.

I'm a developer. There is really no excuse for writing code that is not thread safe. There are so many mechanisms in modern compilers that allow you to prevent threads stepping on each others toes that it boils completely down to poor coding if this is the case.

Another scenario that will cause lock ups is a simple infinite loop, where a condition required to exit the loop is never met and there is no contingency for failing gracefully. Perhaps there is a recursive function that never has a resolution, which will repeat forever adding itself to the stack until the space in the stack is full and lead to a stack overflow and catastrophic failure. It's all pissing in a creek as we're never going to know the details of what's going on.

 

[boutrosinit]

Got the 360 special ed, got a HD too obviously as I'm not retarded.

I drove to Niko's place and now I'm watching Ricky Gervais on the TV. It's his stand-up show! No crashes thus far

 

[Ranger X]

I played for 6 hours tonight. 60gb PS3 here. No freeze.

Still this is the buggiest PS3 game i've played so far. Even more apparent bugs than GTAs on PS2. I've even had a FPS lock-up for 2-3 seconds.

 

[CowboyAstronaut]

I've still yet to experience the freeze and I've played tons of hours on the 360 version.

Has this happened to any falcon 360 owners?

 

[kassatsu]

no problems yet, 60 gb

Same for me. About 3 hrs in.

 

[teh_pwn]

I'm a developer. There is really no excuse for writing code that is not thread safe.

I am too, but in a completely different industry.

My point wasn't that "oh, it's complex, give them a break." It was more like "I'm not defending them, but a fix could take weeks."

 

[Marty Chinn]

Defending? I'm being realistic. Complex multi-threading tasks that very rarely cause crashes on particular models of variations of hardware is not something you can quick fix. Unless Rockstar has some way of getting a bug report that gives them a stack crawl and a line number, this could take a long ass time to fix if it is a race condition.

A race condition is when you have a condition in multi-threading code such that two threads operate out of order depending on variations of conditions or hardware. It's when something that is supposed to rigidly be defined in data flow operates very rarely out of order. It's extremely hard to find because it requires you to start by looking at the entire source. You've got to have something to narrow down something massive like this. It's possible that a race condition could work in testing, but with a slight variation in hardware architecture (revision) that something happen out of order too quickly.

I'm in game development, and I certainly understand what a pain in the ass it is for a developer to debug a situation as I've encountered it myself where there would be issues between a debug and release build and there was very little info to go by to what changed.

However, I am honestly finding it harder to believe now that I know it has been duplicated on a few debug kits and it crashed on me on my devkit. The simple fact that it was that easy for us to find the issue with on development hardware and knowing they must have had a ton of debug kits for the testers to test such a game, you would have thought they would have encountered this at some point.

I don't understand how this could have slipped by so easily between Rockstar, Take-Two, and Sony who all tested it. Somehow it went by overlooked because I don't think Rockstar knew about it. Had they known about it, I would have thought by the time it shipped a patch would be ready. I honestly think they're just hearing this as of launch for the first time and I just don't get how this happened. The only possible explaination is that somehow the production disc somehow causes the problem that a normal send of the code via target manager, or the debugger won't catch as well as a burn on a blu ray-r disc. I don't know what would change from a burned disc to the production disc though.

This really is a mystery and a part of me understands the situation whle the other part of me questions how this went unnoticed.

 

[ComputerNerd]

I'm a developer. There is really no excuse for writing code that is not thread safe. There are so many mechanisms in modern compilers that allow you to prevent threads stepping on each others toes that it boils completely down to poor coding if this is the case.

Another scenario that will cause lock ups is a simple infinite loop, where a condition required to exit the loop is never met and there is no contingency for failing gracefully. Perhaps there is a recursive function that never has a resolution, which will repeat forever adding itself to the stack until the space in the stack is full and lead to a stack overflow and catastrophic failure. It's all pissing in a creek as we're never going to know the details of what's going on.

Yup. First step for Rockstar is to replicate the bug. Preferably on a test unit. Which in and of itself can take a few hours if they have no units in the office that have the bug.

I always hated troubleshooting lockups. Hell, I hate troubleshooting runtime errors in general.

 

[EviLore]

Ok, for those of you who got it working through disabling connecting to PSN, was your crash during the intro cinema or at some point during gameplay?

After someone mentioned a temporary lock-up during PSN related activity earlier in the day, I disabled PSN. Game still crashed within 15 minutes (mine's the intermittent gameplay crash, not the intro crash).

 

[teh_pwn]

I'm in game development, and I certainly understand what a pain in the ass it is for a developer to debug a situation as I've encountered it myself where there would be issues between a debug and release build and there was very little info to go by to what changed.

However, I am honestly finding it harder to believe now that I know it has been duplicated on a few debug kits and it crashed on me on my devkit. The simple fact that it was that easy for us to find the issue with on development hardware and knowing they must have had a ton of debug kits for the testers to test such a game, you would have thought they would have encountered this at some point.

I don't understand how this could have slipped by so easily between Rockstar, Take-Two, and Sony who all tested it. Somehow it went by overlooked because I don't think Rockstar knew about it. Had they known about it, I would have thought by the time it shipped a patch would be ready. I honestly think they're just hearing this as of launch for the first time and I just don't get how this happened. The only possible explaination is that somehow the production disc somehow causes the problem that a normal send of the code via target manager, or the debugger won't catch as well as a burn on a blu ray-r disc. I don't know what would change from a burned disc to the production disc though.

This really is a mystery and a part of me understands the situation whle the other part of me questions how this went unnoticed.

Oh, wow. In that case it looks like something must have slipped through late in development.

That's probably easier to fix, but at the same time that may be indicative of future problems.

 

[Marty Chinn]

Yup. First step for Rockstar is to replicate the bug. Preferably on a test unit. Which in and of itself can take a few hours if they have no units in the office that have the bug.

I always hated troubleshooting lockups. Hell, I hate troubleshooting runtime errors in general.

The problem is it's hard to figure out which PS3 is going to cause it and which is not. A friend ran it on two debug kits that had the same date and appeared to be the same line but only one of them caused the crash.

 

[jetjevons]

Don't get a heart attack, and wait a few days.

That should be R* official PR to the guys that waited X amount of hours in the rain to get the game at the midnight launch events because they foolishly thought it might be exciting to play the game on launch day.

 

[Lord Error]

Welp, I'm happy to report no crashes or weird hiccups of any kind on my 60GB PS3 after about an hour of playing it.

I really have no idea what could be the distinguishing factor here.

 

[ScrabbleBanshee]

Oh, wow. In that case it looks like something must have slipped through late in development.

That's probably easier to fix, but at the same time that may be indicative of future problems.

This is what subversion is for (or cvs, or Team Server or whatever the hell they might be using). Source control for the win... I'm sure the blame trail will point to the poor sap in the end and there will be blood.

Conspiracy theory alert...

Spoiler

corporate sabotage

.

 

[ComputerNerd]

This is what subversion is for (or cvs, or Team Server or whatever the hell they might be using). Source control for the win... I'm sure the blame trail will point to the poor sap in the end and there will be blood.

Conspiracy theory alert...

Spoiler

corporate sabotage

.

haha, I use CVS at my job. It's OK, but annoying.

 

[NinjaCodah]

It is absolutely mind boggling that people are defending Rockstar with regards to this situation in any way, shape, or form.

Look, I don't really want to defend them. I hate it too when things don't work. However, I really do believe it's an issue that even the best QA on the planet could have missed, or it was slipped late. It obviously doesn't occur on all PS3s or 360s. I've been programming game engines and games for most systems for a while now. And when i see the scope of gta4, how big it is, how many mechanics there are and, well, all the tech behind it, i'm simply amazed that this thing even runs. It's incredible. Trust me, there is nothing like it out there. No matter how good looking or fun other games might be, there are none that have as many mechanics or systems as gta does; not even close. No such obvious bug could live against a development team like that, i refuse to believe it. This bug is not an obvious one.

To me, the reason you can't compare it to other games is simply because no other games does nearly as much as gta does!

I hope they fix the problem soon.
PEACE

 

[Animal]

I don't know why some sites are reporting that it is only the 60GB model that is affected. I was playing this evening on a week-old 40GB model and it froze twice during 5 hours of play.

 

[Marty Chinn]

This is what subversion is for (or cvs, or Team Server or whatever the hell they might be using). Source control for the win... I'm sure the blame trail will point to the poor sap in the end and there will be blood.

Conspiracy theory alert...

Spoiler

corporate sabotage

.

Well the problem with source control is it helps if you know about when the bug was introduced. If you don't know when it was, and I doubt they knew about the problem until recently, it doesn't buy you anything.

Source control is great when you can target a general range and start pinpointing it down to what change list caused the new problem. But in this case for all they know, it never worked right to begin with. BTW I bet they use Perforce =)

 

[Moonwalker]

Marty are you able to profile the spu usage? I have a feeling the problem might be related to using all of the spus at full throttle, exposing a processor bug and or faulty spu. Could be why certain machines are experiencing it and others are not. Could also be related to maybe a certain instruction being excessively used causing overheating on parts of the processor and hence failure. I know this has happened before with pc gpus/processors, all it takes is certain code to expose a manufacturing flaw.

 

[Marty Chinn]

Look, I don't really want to defend them. I hate it too when things don't work. However, I really do believe it's an issue that even the best QA on the planet could have missed, or it was slipped late. It obviously doesn't occur on all PS3s or 360s. I've been programming game engines and games for most systems for a while now. And when i see the scope of gta4, how big it is, how many mechanics there are and, well, all the tech behind it, i'm simply amazed that this thing even runs. It's incredible. Trust me, there is nothing like it out there. No matter how good looking or fun other games might be, there are none that have as many mechanics or systems as gta does; not even close. No such obvious bug could live against a development team like that, i refuse to believe it. This bug is not an obvious one.

To me, the reason you can't compare it to other games is simply because no other games does nearly as much as gta does!

I hope they fix the problem soon.
PEACE

I agree with this for the most part. The complexity of the game is just sure amazing and I can only imagine what chaos it is to debug some weird obscure bug. I still find it hard to believe that they didn't discover this given that I'm sure they had a ton of debug kits and between a couple people already their debug kits caught the problem right away. That part I don't get how it was missed.

 

[Raistlin]

Played for about 2.5 hours tonight ... no problems. Brings me to about 4 - 4.5 hours total.




Well, I should detail the above. When I just exited the game, it caused my PS3 to reset. No problems in the game though.