Diet for server processor
There was a server on the Supermicro X5DAL with 2 Intel Xeon 2.8 GHz physical processors in 604 sockets. It hit me in 2008, being a bit second-hand. He worked under 2003 Server, but the trouble was: he was buzzing heavily (like an airplane). Buzzed and warmed up (about 400 watts). It is impossible to slow down the fans (2 on the processors and 2 in the power supply) - it overheats.
As time went on, productivity began to be missed. Replaced it with "self-assembly" on i7. But to throw such a piece of iron was a pity. Why not lower the voltage of the processors? It is known that CPU consumption is proportional to the clock frequency and the square of the supply voltage.
Regular Upit = 1.5V (and it is written on the processor). We ought to lower it. In the BIOS, there is no such option, which means that it is necessary to take on the soldering iron.
')
The power controller for both processors is single, 4-phase (4 chokes are visible in the photo), HIP6311. The processor has 5 special outputs with the names VID0-VID5, some are shorted to ground (GND) (correspond to log.0), some are hanging in the air (correspond to log.1). This code is supplied by five separate wires from one of the processors to the HIP6311 terminals of the same name, and the latter stabilizes the desired supply voltage at the output of the circuit. There is a match table in the datasheet.
The table goes on and on, but this part is still sufficient.
The source is powerful enough: each processor consumes about 90 W, respectively, the source provides ~ 200 W, and at 1.5 V it is 133 A.
For testing, a single processor is enough. The radiator just lay on top of the thermal grease (the original fasteners are reliable and strong, but it is laborious to put it on and off every time).
Soldering off the legs of the HIP6311 chip, I closed some on GND, choosing less than the specified supply voltage, but minimizing the number of rations. Naturally, all switching and soldering was done on the off system.
Otpaivaya next leg, I hear: sparks. It turns out that the soldering iron was grounded, but the power supply was turned on through a homemade power meter, then without a ground contact. 110VAC from the power supply capacitors sparred to the soldering iron through the HIP6311 leg. Clearly, the microcircuit could not stand the electric shock. The processor before soldering removed - well.
Again, throw out a pity. Thought-thought ... Ebay! Ordered, come, soldered. He bent the legs, extended the wires - everything became easier. It was possible to attach a dip-switch, but so it happened. As a result, the CPU earned 1.15V.
For indication used 3-digit mini-voltmeter. In the standard configuration, it measured 0.0-33.0V, accuracy of 0.1V is not enough. He removed the lower divider resistor (by sketching the circuit and estimating the divider coefficient) - it turned out to be a voltmeter 0.00-3.30V, however, the point cannot be changed without flashing. 1.15V displayed as 11.5, but it is not so bad. (it was possible, of course, to solder the tester, but then I also used a voltmeter to repair AGP video cards).
The processors started up, worked stably and warmed less. Removed + power supply of fans from the power supply unit, connected with + power supply of the processor fans.
I thought to make for them a separate adjustable pulse stabilizer on the LM2596, but it turned out to be enough just to apply + 5V to them. Control of heating under continuous load - on blow +40 degrees. Fine! The noise became quiet (and with a pleasant timbre, but this, of course, an amateur).
Disappointments 2:
1) processors do not support hardware virtualization. And I wanted to make a virtual machine server ...
2) Windows XP Home, the sticker of which was on the case, supports only 1 physical processor. Found out after activation.
The video card was installed by the GF5500 (it had a stabilizer burned, replaced it with a KIS-3R33S with a voltage change of up to 1.2V). Not very new toys coming.
For a meaningful activity put CentOS 6.5.
Clearly, before starting such a server in production, you should think for a while, but there are no particular problems at first glance (except for the need to test the reliability and the fact of intervention by the soldering iron on the motherboard): most importantly, the thermal and electrical load on all components has decreased. The voltage is reduced almost in the regular way. In modern matplata processor voltage can be adjusted from the BIOS, increasing for overclocking, for example, and reducing to save energy and reduce noise.
There are not enough figures for power consumption, but the meter is currently reworking (I connect the USB interface and analog sensors for datalogging).
www.intersil.com/content/dam/Intersil/documents/hip6/hip6311.pdf
As time went on, productivity began to be missed. Replaced it with "self-assembly" on i7. But to throw such a piece of iron was a pity. Why not lower the voltage of the processors? It is known that CPU consumption is proportional to the clock frequency and the square of the supply voltage.
Regular Upit = 1.5V (and it is written on the processor). We ought to lower it. In the BIOS, there is no such option, which means that it is necessary to take on the soldering iron.
')
The power controller for both processors is single, 4-phase (4 chokes are visible in the photo), HIP6311. The processor has 5 special outputs with the names VID0-VID5, some are shorted to ground (GND) (correspond to log.0), some are hanging in the air (correspond to log.1). This code is supplied by five separate wires from one of the processors to the HIP6311 terminals of the same name, and the latter stabilizes the desired supply voltage at the output of the circuit. There is a match table in the datasheet.
The table goes on and on, but this part is still sufficient.
The source is powerful enough: each processor consumes about 90 W, respectively, the source provides ~ 200 W, and at 1.5 V it is 133 A.
For testing, a single processor is enough. The radiator just lay on top of the thermal grease (the original fasteners are reliable and strong, but it is laborious to put it on and off every time).
Soldering off the legs of the HIP6311 chip, I closed some on GND, choosing less than the specified supply voltage, but minimizing the number of rations. Naturally, all switching and soldering was done on the off system.
Otpaivaya next leg, I hear: sparks. It turns out that the soldering iron was grounded, but the power supply was turned on through a homemade power meter, then without a ground contact. 110VAC from the power supply capacitors sparred to the soldering iron through the HIP6311 leg. Clearly, the microcircuit could not stand the electric shock. The processor before soldering removed - well.
Again, throw out a pity. Thought-thought ... Ebay! Ordered, come, soldered. He bent the legs, extended the wires - everything became easier. It was possible to attach a dip-switch, but so it happened. As a result, the CPU earned 1.15V.
For indication used 3-digit mini-voltmeter. In the standard configuration, it measured 0.0-33.0V, accuracy of 0.1V is not enough. He removed the lower divider resistor (by sketching the circuit and estimating the divider coefficient) - it turned out to be a voltmeter 0.00-3.30V, however, the point cannot be changed without flashing. 1.15V displayed as 11.5, but it is not so bad. (it was possible, of course, to solder the tester, but then I also used a voltmeter to repair AGP video cards).
The processors started up, worked stably and warmed less. Removed + power supply of fans from the power supply unit, connected with + power supply of the processor fans.
I thought to make for them a separate adjustable pulse stabilizer on the LM2596, but it turned out to be enough just to apply + 5V to them. Control of heating under continuous load - on blow +40 degrees. Fine! The noise became quiet (and with a pleasant timbre, but this, of course, an amateur).
Disappointments 2:
1) processors do not support hardware virtualization. And I wanted to make a virtual machine server ...
2) Windows XP Home, the sticker of which was on the case, supports only 1 physical processor. Found out after activation.
The video card was installed by the GF5500 (it had a stabilizer burned, replaced it with a KIS-3R33S with a voltage change of up to 1.2V). Not very new toys coming.
For a meaningful activity put CentOS 6.5.
Clearly, before starting such a server in production, you should think for a while, but there are no particular problems at first glance (except for the need to test the reliability and the fact of intervention by the soldering iron on the motherboard): most importantly, the thermal and electrical load on all components has decreased. The voltage is reduced almost in the regular way. In modern matplata processor voltage can be adjusted from the BIOS, increasing for overclocking, for example, and reducing to save energy and reduce noise.
There are not enough figures for power consumption, but the meter is currently reworking (I connect the USB interface and analog sensors for datalogging).
www.intersil.com/content/dam/Intersil/documents/hip6/hip6311.pdf
Source: https://habr.com/ru/post/218469/
All Articles