Slashdot Mirror
Boot Log Messages On A Pre-Production Processor
Check this Boot log. This is a log from the Power4 Processor which is in the pre-production stages, and since it's in the pre-production stages, they removed the BOGOMIPS value. Nevertheless - it looks very interesting. Keep up the good work, IBM!
Answered in this comment.
"If one is really a superior person, the fact is likely to leak out without too much assistance" -- John Andrew Holmes
That's one thery.
Another thery is making one CPU capable or running instrucitons from more then one stream in the same clock cycle and allowing it to choose which functional units get assigned to which thread on every clock will be faster. The Alpha 21364 designers obviously beleve in this thery, since that is how it works.
A simple thought exparament will show the multiple threads CPU is faster if it has all the functional units the multiple CPU on a die version has. Of corse if it is far simpler to put a whole bunch of CPUs on one die then to make a whole bunch of FUs for one CPU, then it might not be such a win.
Since both CPUs are due out next yearish I guess we can find out who is right then.
Personally I think the Alpha approch will win for the more common workloads we see today. But that the multi-CPU method will be far simpler to ramp up, and will win on more losely cuppled problems.
These porting people should know that bogomips does not relate to the processor's performance.
The BogoMIPS value is a measure of how many million times per second a processor can perform "do nothing" executions. It's measured at startup and used to tweak some timing loops in the kernel. From a marketing standpoint, however, there is absolutely nothing sexxy about publishing how good your cpu is at doing nothing.
But both Bogomips and a boot log can hardly show the real advantages
since it's in the pre-production stages, they removed the BOGOMIPS value.
Obviously plenty of people think the bogomips value is SO FSCKING IMPORTANT that they felt it necessary to hide it and then claim that it has no meaning? THis is like the gov't claming there's no base near groom lake, Nevada, and infact that there's nothing there at all, yet threatens to arrest/detain/kill (signs say "use of deadly force authorized") anyone who tries to go there.
Get over yourselves. Bogomips matter. Your own paranoia over them just proves it all the more.
2.95.2 19991024 (release/franzo)) #147 Tue May 30 15:54:28 CDT 2000
Took 147 compiles to get it to work?
Yes, some interesting information about the CPU can be calculated from the BogoMIPS value. But there is one really important reason they matter: people think they do.
it's green.
The rule of thumb that's always worked for me is that each CPU added increases performance by something around 60% of the last increase. So the first CPU adds a 100% overall boost (from nothing to doing something), the second 60% overall, the third 36% overall, and so on. Doesn't always work, but it helps you get a rough idea of what's going on.
it's green.
You are right, this isn't a PowerPC chip; MacOS might not run, and I agree there would probably be a Linux version first. But if for some reason I had to choose, I'd go with the MacOS X.
it's green.
This is not a SOFTWARE development box. So, really, there is most likely no need for all those disks. Unless they are testing the IO or something.
--
"What beats rock?" "Nuthin' beats rock!"
Four-digit slashdot ID. Recognize.
Can I post my linux bootlog for my TI-86 caculator? Huh? huh?
duh!
odds are that most of that space that was not able to be mapped into a partition table (read: MOST of that dasd) is jfs file space dedicated to an AIX installation or two.
Yes, but this is not a production unit, it's a development machine used to get Linux running on the new CPU. It makes little sense to me to have that many disks. Even if you were testing RAID you'd only need about 5 or 6 to do some nice tests. I have a SUN E450 at work that has a bunch of disks in RAID5 also, but it's also a production box...
The real problem with on die SMP is that it wastes 'resources'. Most of the next generation of 'RISC' processors will be SMT (symmetric multithreaded) instead which allows them to better utilize of the available transistor count. Expect to see the competitors with 4x or 8x SMT CPU's.
The other advantage of SMT is that it allows your cpu to tolerate higher memory latencies because it isn't such a big deal to stall a thread (waiting for memory) because the other threads will continue to utilize any functional units they need.
What I would like to know is, where is the boot message bringing the second CPU online? It looks like they are probably just running one CPU in this log.
see, that was my thought, too... I checked the decimal places a few times.
/. linux tuning forum. Post your bootlog, and see if everybody else can pull it apart... kinda like 100,000 mechanics all trying to work on the same car at the same time, but most of them have only worked on GM vehicles, and this is a foreign model...
Maybe we should have a
"It's tough to be bilingual when you get hit in the head."
Over the years, I've found it is either half, the same, or double the CPU speed. I don't know much more than that. :)
How about dmesg | less in an eterm running on my 1280x1024 desktop? Besides, I wouldn't be able to read it even at 1024x768 frame buffer mode, as it takes my system 30 seconds to boot. You blink and the screen is cleared and giving you a login prompt :)
IBM doesn't control the development of Linux, and doesn't know where Linux is going to be in a year's time when the power4 finally gets released. (like anybody does!). Their primary interest is probably to make sure that the kernel doesn't hork on some unexpected change in the processor design.
AIX, on the other hand, is IBM proprietary. If Big Blue doesn't have AIX running on that CPU by the time it comes out the door, their ass is gonna be grass. They don't have a legion of Open Source hackers to do their dirty work for them.
Then there's the AS/400 which is also capable of running on the PowerPC CPU.
Put a handfull of applications, multiple development sources and enough disks to test the I/O on each of those OS's and that stack of disks starts to look conservative.
--
Free Software: Like love, it grows best when given away.
I can see it now, "Buy a new PV, see how fast your computer does nothing!"
That almost looks like it was booting on a test mule Powermac 9600 (with a P4 proc). Yellowdog linux offers the champion server for Macs, and there's that funny line about Macintosh NVram mapping or something. Also, the FD controller is the same one used on late model 9600s. Hmmm....
-dave
This is not a sig. this is a duck. quack.
Remeber, the POWER4 is "NOT" a PC processor. It's target platform is high-end RS/6000 machines. Think "big iron". I'll bet that's a mid-range configuration.
I take offense to this comment. MacOS is a fine operating system, as is Linux. If you are the type of person that likes to do things for themselves then that's ok, but some of us don't have time to learn the entire language to work under Linux. Now, don't get me wrong, Linux is a powerful operating system and I may be a little biased because I have been a mac user all my life but they are both operating systems with their own advantages and disadvanteges. Although I despise microsoft because of their tactics and not necesarily stealing of ideas from apple, but trying to pass them off on their own...Sorry, I went off on a tangent. MacOS and Linux, and Unix, and Unix, and whatnot are trailing behind Windoze and their assorted OS'es in software, but we all still in need of a common bond to bring us together to fight the evil of William Gates III!!! HAHAHAHAHAHAHAHAHAHAHAHAHAHAHA.
A persons true power is a combination of mind and body. To reach full potental you must get both to work as a si
Linux version 2.3.99-pre9 (root@enigma) (gcc version egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)) #1 Sun Jun 4 01:17:19 PDT 20000 .html
:) ROAR! Also, my Mhz/BogoMIPS aren't top secret ;)
BIOS-provided physical RAM map:
e820: 000000000009fc00 @ 0000000000000000 (usable)
e820: 0000000000000400 @ 000000000009fc00 (reserved)
e820: 0000000000010000 @ 00000000000f0000 (reserved)
e820: 000000000fef0000 @ 0000000000100000 (usable)
e820: 0000000000008000 @ 000000000fff0000 (ACPI data)
e820: 0000000000008000 @ 000000000fff8000 (ACPI NVS)
e820: 0000000000010000 @ 00000000ffff0000 (reserved)
user-defined physical RAM map:
e820: 000000000009f000 @ 0000000000000000 (usable)
e820: 000000000ff00000 @ 0000000000100000 (usable)
On node 0 totalpages: 65536
zone(0): 4096 pages.
zone(1): 61440 pages.
zone(2): 0 pages.
Initializing CPU#0
Detected 654593298 Hz processor.
Console: colour VGA+ 80x25
Calibrating delay loop... 1304.17 BogoMIPS
Memory: 254828k/262144k available (1614k kernel code, 6928k reserved, 122k data, 204k init, 0k highmem)
Dentry-cache hash table entries: 32768 (order: 6, 262144 bytes)
Buffer-cache hash table entries: 16384 (order: 4, 65536 bytes)
Page-cache hash table entries: 65536 (order: 6, 262144 bytes)
kmem_create: Poisoning requested, but con given - bdev_cache
Inode-cache hash table entries: 16384 (order: 5, 131072 bytes)
kmem_create: Poisoning requested, but con given - inode_cache
CPU: L1 I Cache: 64K L1 D Cache: 64K
CPU: L2 Cache: 512K
CPU: AMD AMD-K7(tm) Processor stepping 02
Checking 386/387 coupling... OK, FPU using exception 16 error reporting.
Checking 'hlt' instruction... OK.
POSIX conformance testing by UNIFIX
mtrr: v1.36 (20000221) Richard Gooch (rgooch@atnf.csiro.au)
PCI: PCI BIOS revision 2.10 entry at 0xfda01, last bus=1
PCI: Using configuration type 1
PCI: Probing PCI hardware
PCI: Using IRQ router VIA [1106/0586] at 00:04.0
isapnp: Scanning for Pnp cards...
isapnp: Calling quirk for 01:00
ISAPnP: SB audio device quirk - increasing port range
isapnp: Calling quirk for 01:02
isapnp: AWE32 quirk - adding two ports
isapnp: Card 'Creative SB AWE64 PnP'
isapnp: 1 Plug & Play card detected total
Linux NET4.0 for Linux 2.3
Based upon Swansea University Computer Society NET3.039
kmem_create: Poisoning requested, but con given - skbuff_head_cache
NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
NET4: Linux TCP/IP 1.0 for NET4.0
IP Protocols: ICMP, UDP, TCP
IP: routing cache hash table of 2048 buckets, 16Kbytes
TCP: Hash tables configured (established 16384 bind 16384)
ACPI: "AMI" found at 0x000fb080
ACPI: found platform errata 0x00000030
ACPI: unreserved table memory @ 0x0fff00b0!
Starting kswapd v1.6
fb: Voodoo3 memory = 16384K
fb: MTRR's turned on
tdfxfb: reserving 1024 bytes for the hwcursor at 0xd1803000
Console: switching to colour frame buffer device 80x30
fb0: 3Dfx Voodoo3 frame buffer device
pty: 256 Unix98 ptys configured
Uniform Multi-Platform E-IDE driver Revision: 6.30
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
VP_IDE: IDE controller on PCI bus 00 dev 21
VP_IDE: chipset revision 6
VP_IDE: not 100% native mode: will probe irqs later
hda: WDC AC418000D, ATA DISK drive
hdc: CD-532E, ATAPI CDROM drive
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
ide1 at 0x170-0x177,0x376 on irq 15
hda: 35239680 sectors (18043 MB) w/1966KiB Cache, CHS=2193/255/63
hdc: ATAPI 32X CD-ROM drive, 128kB Cache
Uniform CD-ROM driver Revision: 3.08
Partition check:
hda: hda1 hda2 hda3
Floppy drive(s): fd0 is 1.44M
FDC 0 is a post-1991 82077
Serial driver version 4.93 (2000-03-20) with MANY_PORTS SHARE_IRQ SERIAL_PCI ISAPNP enabled
eepro100.c:v1.09j-t 9/29/99 Donald Becker http://cesdis.gsfc.nasa.gov/linux/drivers/eepro10
eepro100.c: $Revision: 1.29 $ 2000/03/30 Modified by Andrey V. Savochkin and others
eth0: OEM i82557/i82558 10/100 Ethernet, 00:80:5F:F7:E9:27, IRQ 9.
Board assembly 692290-002, Physical connectors present: RJ45
Primary interface chip i82555 PHY #1.
General self-test: passed.
Serial sub-system self-test: passed.
Internal registers self-test: passed.
ROM checksum self-test: passed (0x24c9f043).
Receiver lock-up workaround activated.
[drm] Initialized tdfx 0.0.1 19991009 on minor 63
ISAPnP reports AWE64 WaveTable at i/o 0x620
Soundblaster audio driver Copyright (C) by Hannu Savolainen 1993-1996
sb: Creative SB AWE64 PnP detected
sb: ISAPnP reports 'Creative SB AWE64 PnP' at i/o 0x220, irq 5, dma 1, 5
sb: 1 Soundblaster PnP card(s) found.
kmem_create: Forcing size word alignment - nfs_fh
VFS: Mounted root (ext2 filesystem) readonly.
Freeing unused kernel memory: 204k freed
Adding Swap: 265064k swap-space (priority -1)
[EXT II FS 0.5b, 95/08/09, bs=4096, fs=4096, gc=55, bpg=32768, ipg=32320, mo=ffffffea]
Note the badass 2.3.99-pre9 isapnp
Marketing done by IBM is a weird thing. I have no clue what you do or where you do it...but IBM doesn't go much for bragging rights on this kind of stuff. Its more like...let's sell it to the people who use our products and optimize our products for it. I will say this for IBM...as a Big Blue employee but a much longer Linux lover...IBM is really going with this Linux thing. I almost feel as if it being pushed more than AIX. Course..being the Linux guy in my area...I see more of Linux than anything.
So heck...if they are bringing this to their big buyers and saying..."Use AIX or Linux"..AWESOME! At least they give the option.
That was a dumb way to put it...but what I meant is the recent Linux movement in big corporations. Had to clarify..the way I said it sounded really bad.
IMO, a devel box needs more space than a production box (especially in a team environment).
Bill - aka taniwha
--
Leave others their otherness. -- Aratak
lessse.. post your opinions, your flamebate. Follow the masses your cool.
lahooosers
Its IBM! Hard drives grow on trees at IBM.
Well, under AIX, under which the power4 chip will likely run, you can pull the plug, and it will replay its journal. Also, on this kinda iron (H&J 50s and above) you usually have your disks in a separate case running ssa. You can shut down the CPU without losing much data because (at least in my company's case) the machines are configged to flush ~1sec. Of course this is a 8-proc machine with power to spare for such things.
Lowmag.net
Without performance indicators, what the hell does this log show us?
Sure it shows IBM supporting linux, but isn't that the fad for everything else? I mean i find it more intriguing to see linux running on webpads, palm pcs and miniature servers more so then another 'look what linux can run under'
OTH, go blue go.. i want a Power4 running MacOS X at 2 ghz.
hehe.. oops, i said i wanted MacOS X running under one of these.. oh well, linux would still be cool
swwwwwweeet revalation
sweeeeeeet surrenduring
Thinking of you.. thinking of you
oops.. didn't mean to trail off like that, guess there isn't much else to speak about on this subject :)
- Bootlog Battle ('d00d' my bootlog consumes its own 1Gb disk')
- Uptime Follies ('What do you mean it's impossible? This machine HAS been up for 23.723243 years!!!')
- Peripheral Cramming ('It was tough making PCI-slot splitter cables so I could cram those seven Geforce cards in there, but you should see the results!')
- Megahertz Madness ('I can melt the case on my overclocked Celeron in 2.5 seconds.. and that's MOLTEN baby!')
More to come, I'm sure...-'fester
raid5: measuring checksumming speed
8regs : 48091.100 MB/sec
32regs : 93871.305 MB/sec
using fastest function: 8regs (48.387 MB/sec)
Isn't 93871.305 > 48091.100?
My bootlong Starting Windows 98.... :)
I can't believe I bit... *groan*
And so it begins...
-'fester
Apparently some people missed line 20.
Well I can see this configuration happening. One of the sun servers used by a cs prof (david beazley of python fame) at my school has a similiar setup. Two drives in a RAID 0 configuration for the system disk and 10 drives in a RAID 5 for the rest of the space usuage. You get fast performance and good fault tolerance.
"When you sit with a nice girl for two hours, it seems like two minutes. When you sit on a hot stove for two minutes, it
It wasn't an attack on MacOS. This is "Heavy Iron" running Power4, not some PowerPC chip. I bet AIX will be the only UNIX capable of using that hardware to its max. Plus, MacOS really doesn't take advantage of multiprocessing (up to OS X it was hardly multitasking, but rather multiprogramming).
The technical capabilities and ease of use of an operating system are still two totally distinct things right now.
I bet that MacOS couldn't run on something like this anyways unless they did a custom hack.
First of all... Why the hell would you want to run Mac OS X on the thing? Actually, I have to say the same about Linux unless it's drastically improved in the next year.
Hrmm... I don't know. Let me think. Could it be because we use those systems and the chip is FAST???? It's not about the marketing. It's about the raw power.
Besides, the only thing AIX is good for is giving sysadmins nightmares and giving your development team ulcers when they try to compile their code on the newest version of the OS. I swear. IBM changes their damn networking code's interfaces and function parameter data types so often you'd think they're trying to drive their customer base to suicide. There's a reason I pronounce is "Aches" sometimes. Oh, and don't get me started about SMIT and the seemingly deliberate effort to break shell scripts that run just fine across other systems by using non-standard command arguments.
Of course, IBM's attitude has always been that the standard is whatever IBM chooses it to be. Witness SNA, Microchannel, and their late forced entries into the microcomputer and personal computer markets.
If it's for-profit but free, you're not the customer -- you're the product (e.g., the Slashdot Beta's "audience").
repeat this to yourself: BOGOMIPS ARE MEANINGLESS
they are in NO way a measure of preformance. it is nothing more than an empty delay loop.
huh? /proc/cpuinfo
How?
My Athelon 700 does only 696.32 bogomips according to
--
If Microsoft is the solution, I want my problems back
they both shop with ReiserFS.I have been running
ReiserFS on 3 boxes and no problem what so ever.
42
Is this similar to the DOS 'No keyboard present. Press F1 to continue'??
Ya think they really needed all those SCSI disks just for getting Linux to boot on a Power4 CPU? I mean cripes, they have like 15 17GB disks and 2 8.5GB disks in there! I can see a few for testing RAID, but 17 disks?!? Isn't that a little overkill?
The BogoMIPS value is a measure of how many million times per second a processor can perform "do nothing" executions. It's measured at startup and used to tweak some timing loops in the kernel. From a marketing standpoint, however, there is absolutely nothing sexxy about publishing how good your cpu is at doing nothing.
"If one is really a superior person, the fact is likely to leak out without too much assistance" -- John Andrew Holmes
or there's this bit of drivel I picked up a couple years ago....
...
Hey, propeller heads! Have you got Unix DOWN COLD? Recompiled the kernel so many times you've got it menued and scripted? Setting speed records for setting up sendmail correctly in less than four months? Then, have we got a challenge for you! Yes, it's AIX, from those wonderful people who brought you OS/7 Ferrengi! Take the most obscure and difficult parts from both SVR4 and BSD, add in more bugs than an African termite mound, and you've got an opperating system that GUARANTEES your job. Yes, now you can have everyone in your organization frustrated and furious, and only YOU know how to make it work! More obscure commands than Novell Netware 8.8! More bloated C++ code than MSWindows 21st Century! So call the IBM division of Walt Disney Enterprises TODAY! And make sure that no one at your place of work can AFFORD to fire you! And remember, even though we work for him, WE'RE not Mickey Mouse! So look for the nerd in the blue suit and ask for AIX. You'll be glad you did!
Oh, John, I just LOVE a man who knows AIX
(Don't blame me, I didn't do it.)
Any sufficiently advanced civilization is indistinguishable from Gods.
It's ridiculous that Linux does not have a journaling filesystem, yet.
Actually, it has 3. ReiserFS is the most stable at this time IMHO, but xfs and ext3 are shaping up nicely. They are not yet in the main tree, but are easy to add in. I'm pretty sure that Suse ships with Reiser.
would a boot log be posted as a story :)
----------
"They misunderestimated me." --George W Bush, Nov. 6, 2000
"The new wave is not value-added; it's garbage-subtracted" - Esther Dyson, Dec 1994
The people who ported Linux to this box may fully realize that, but they also fully realize that many people outside their small circle will start making all sorts of wacky pronouncements based on this Meaningless Indicator of Processor Speed.
--Joe--
Program Intellivision!
...and IBM probably doesn't want that.
For example, if someone figured out that this pre-production chip was only running at a measly 1GHz, the next day there'd be a huge story on news.com about how the chip was "below expectations" or some other BS.
Starting console mouse services: (no mouse is configured) That's some seriously funny shit.
-Reid
If you thought that boot log was rad, you should see the swap file!
Gamingmuseum.com: Give your 3D accelerator a rest.
How many more years of "Bogomips are not a benchmark" does it take to get to the center of one's brain? BTW, my Athlon 550 does 1104.28 bogomips. :)
And even if it weren't a joke, the original post was *about a bootlog*. I personally don't care about the bootlog; I think it was dumb to post the original story (but you know what opinions are like...everyone's got one. :^)
So, moderators: how is posting a bootlog on a story about bootlogs off topic?
Stating on Slashdot that I like cheese since 1997.