Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?
DEAL: For $25 - Add A Second Phone Number To Your Smartphone for life! Use promo code SLASHDOT25. Also, Slashdot's Facebook page has a chat bot now. Message it for stories and more. Check out the new SourceForge HTML5 Internet speed test! ×

Comment Re:Just in time. (Score 3, Informative) 219

Don't use RAID5 with drives over 1TB.

a) a RAID5 rebuild takes many hours, b/c it involves reading the entire disc.
b) drives from the same production batch tend to cluster failures.
c) I recall reading that the uncorrectable read error rate tends towards the 2TB mark.

That is, chances are very good that a single drive failure will become a 2-drive failure during a rebuild.

RAID6 or nothing.

Comment Re:Air resistance. (Score 1) 1184

Probably, Northern California, SF Bay. I've lived here for 5 years, and one of the oddest things is that if you get out of the Bay the speed limits on I-5 & I-80 are 70MPH. But once you're in the Bay (I-680, I-880, I-580, I-280, US-101, CA-237, CA-85) speed limits are 55 or 65.

Assuming you CAN go that fast, and assuming that the others in your lane are actually willing to go that fast. For which nobody steps on it in the entrance ramps. So your entrance speed to the highway is likely 40. Maybe 50.

I'm so glad to be moving out of here at the end of the week.

Comment Re:Corporate Personhood, worst conspiracy ever (Score 1) 158

In order to do so, might one also need to make the quorum shareholders of a publically traded corp each plaintiffs for any lawsuit that a corp wants to file ?

Limited personhood perhaps makes sense, but I'd call it merely a 'legal entity' that is empowered to do certain things (financial and legal transactions) and not do other things (vote, unrestricted speech, etc). But that is more complicated than just calling them persons en toto.

Comment Re:Get ready for a new wave of poorly coded softwa (Score 1) 133

For as much as anecdote != data, I took a recent (installed it Thanksgiving 2010) OCZ MLC SSD (30G) SSD, and gave it a database workload wherein a commit was made every 5 seconds. Wore it out in a year. Replaced it in August.

Relevant SMART data (cleaned up so /. doesn't hate on it)
205 Max_PE_Count_Spec       0x0000   -   -   -    Old_age   Offline      -       10000
206 Min_Erase_Count         0x0000   -   -   -    Old_age   Offline      -       8948
207 Max_Erase_Count         0x0000   -   -   -    Old_age   Offline      -       9930
208 Average_Erase_Count     0x0000   -   -   -    Old_age   Offline      -       9655
209 Remaining_Lifetime_Perc 0x0000   -   -   -    Old_age   Offline      -       4

You can't tell me it can't be done with modern SSDs. It merely isn't the case for MOST consumers. But when has Slashdot been made up of MOST consumers?

Comment Re:register starvation (Score 1) 594

It seems that the intent there is to make C into something a lot like C++/STL. Once you've done it once, nothing to prevent Complex types and others being builtins... but this seems to bring C further from its "portable ASM" status.

Not that I'm entirely sure this would be bad... and the struct could be just a typedef. Now all we need is operator overloading, which is also a C++ism. Maybe a C++ minus the type-strictness, and minus all of the automatic copying. and the templates.

Comment Re:register starvation (Score 1) 594

I've seen this done in serialization formats, but never seen it actually done in a program.

So you're saying you want native-size-word (int) followed by the string chars, so you'd malloc(strlen+4) for the dynamic string, and then in the reader read the first 4 bytes and set the ptr to stringPtr=basePtr+(void *)(sizeof(size_t)). It just doesn't feel natural in C as we know it. But then again, I'm sure they could have defined it to be natural. it would be an implied struct.

Comment Re:register starvation (Score 1) 594

I meant that if you were to actually pass the struct as not a structptr, but as the 2 values themselves. After all, you could optimize for fewer regs, or fewer memory accesses. Pointer indirection used to be considered a problem. Albeit not anymore.

Yes, that would be an ABI change, but this was back when they could choose the ABI as they wished... so we can throw many current assumptions of ABI out the window.

Comment Re:register starvation (Score 1) 594

I'll admit to not being an assembly programmer, but for one, I was referring more to what they would have done back then... when register renaming wasn't available. Nowadays many programmers don't care about cycle counting... I know I haven't had to. Writing in perl or C++/STL leaves me too far away from the metal to even know how many cycles I'm wasting.

But since the question is one of yesteryear, lack of regs could be a question. And I was under the impression you weren't supposed to use esi and edi as tmp stores or as the output of an expr. They're ptrs into stack and data-area, respectively.

Comment register starvation (Score 1) 594

The real problem with the addr+len approach is that now every string becomes a struct, or a structptr.

This means that when passing a string to a function, either the string takes up two register/stack slots, or you're passing around a const-ptr (but the contents of the struct are not const), which means one more memory access due to pointer indirection.

x86 and the PDP-11 are register-starved. the x86 has 8 registers, with 4 or 5 available as general-purpose registers.
The PDP-11 was similar with 8 registers total as well.

Slashdot Top Deals

Like punning, programming is a play on words.