IBM and HP have reignited the debate over blade-server power consumption. 

The war of words heated up again after HP released a study claiming that its BladeSystem c-Class servers use as much as 27 per cent less power than IBM's competing BladeCenter-H models.

The study was commissioned by HP and carried out by Sine Nomine Associates. 

Power consumption is a major selling point for blade servers, which are often purchased for their small size and relatively low running costs.

A spokesperson for IBM told vnunet.com that the company is still reviewing the study, but said that early indications suggest that the IBM models being tested were not comparable to the HP machines. 

"If they did a like-to-like comparison under their rules and conditions at best it would be a tie," said the spokesperson.

The two companies were engaged in a debate over testing methods in November, after IBM released a study claiming that its blades used 30 per cent less power than HP's servers.

At the time, HP made a similar claim to IBM, insisting that the tests compared unequal machines.

"Based on what we know about IBM's ageing architecture, we feel strongly that these numbers are not a true apples-to-apples comparison," HP said in response to the IBM study in November.

An IBM spokesperson told vnunet.com that the company challenged HP to a "bake-off" in November, but HP declined. That accusation drew the ire of HP and raised questions of a renewed blade centre spat.

"Show us that challenge. It was not in their release and we never saw it," an HP spokesperson told vnunet.com.

"We honestly do not think a report was ever created and we challenge IBM to share their 'study' from November. We are happy to put C-Class up against BladeCenter-H in a third-party run, apples-to-apples bake off."