Yes, but there are still plenty of "experienced coleagues" who still trundle out these old chestnuts - suggesting that the mere act of writing in assembler will inherently give you better, tighter code.

The point is that there is nothing magic about assembler: it requires skill and experience to write good, tight assembler - just as it requires skill and experience to write good, tight 'C'.

As you gain experience with 'C', you will learn what constructs give good, tight code and what incurs large overheads. Initially, it would be worth your while looking at the generated assembler, and experimenting with your 'C' style to see how it affects the result.
eg, try a simple loop, compile it with Small and then Large Memory models, and compare the generated code!

If you're using Keil, there is a whole section in the C51 Manual on writing 'C' in an optimum way for an 8051.

The IAR article, "Getting the Least Out of Your C Compiler" is well worth a read (it is not 8051-specific):
http://www.iar.com/FilesPubl..._Least.pdf
The following is from the foreword:
"Many established truths and tricks are invalidated when you use a modern compiler. We will demonstrate some of the more common mistakes and how to avoid them, and give a catalog of good coding techniques. The presentation will be illustrated by snippets of real-world code to demonstrate important concepts. An important conclusion is that code that is easy for a human to understand is usually also compiler friendly, contrary to hacker tradition." (my emphasis)

There have been plenty of previous discussions both here and on the Keil site about optimising 'C' for the 8051; eg,
http://www.8052.com/forum/read.phtml?id=77399
http://www.8052.com/forum/read.phtml?id=71812
http://www.keil.com/discuss/