So, maybe I would not be a good (typical) example, since I took part in some physics competitions in a high school (IPhO etc.), and "got trained" in those little tricks (which you can find in a GRE test too) long time ago.
Personally, for PGRE I just needed to refresh agility a little bit, and memorize the important formulae. It took me a week to do so.
I would recommend you one very good Russian book, which I used in a high school as the main material for preparing for (inter)national competitions.
It's name is I.E Irodov's "Problems in General Physics" or shortly "Irodov" (Иродов).
I got the Russian version
, but english one also exists ofc.
The problems are much more difficult then those in the GRE test, but the benefit is great and guaranteed.
My opinion: In order to make sure to get 950+ points,
you need not to practice problems of difficulty like those in pgre, but much more difficult problems instead. I think it's the biggest mistake people do: preparing by solving problems of pgre-level-difficulty, so they end up with failing to finish the test in those 2h 50min. I personaly saw only one pgre test, that one which they send us as an example, it was useful for getting familiar with the content, not as a
practice for the test. The real practice should be on a higher level, which will provide you with cunning and tricks, and that's what gives you speed for pgre problems which are easier (yet tricy), time is essential for this test.
My advice: Take few months, solve 300 various problems from Irodov (cca 70 in mechanics, 50 in electrostatics/magnetism, 40 in thermodynamics...), that book will boost your understanding of general physics, and teach you concepts, usefull time-saving shortcuts and tricks. If you do so, I bet you'll earn at least 940 pts, if you got 840 now. The Irodov problems are quite challenging (still not too difficult, I used the book in a high school!), but it's worthy. Irodov can cover about 80% of the gre content, for the rest 20% (particles, atomic, experimental/statistics, microelectronics...), some additional literature is needed. To put things in a nutshell, there's no magic stick, just work hard with a good book.
Soviets had had very good school science programs and math/phys books, it is shame not to take advance of it.
Some feedbacks from US guys on book I recommended:
http://www.physicsforums.com/archive/in ... 31410.html
http://www.physicsforums.com/archive/in ... 98599.html